Method for improving digital signal processing in radio frequency systems, and system and data carrier signal using same

Abstract

The present invention relates to a method for improving digital signal processing in a radio frequency system (1), in particular in a radio frequency power supply system (1), comprising: providing (101) at least one input signal (210) preferably related to a radio frequency signal of the radio frequency system (1); executing (310) a processing procedure (102), in which the at least one input signal (210) is iteratively processed by a programmable circuit (10) using at least one configurable parameter of the processing, the configuration of the at least one parameter being changed for each processing of the at least one input signal (210) to obtain a respective processing result (220); determining (103) at least one parameter result (240) for optimization of the configuration based on the processing result (220); and providing (104) the at least one determined parameter result for improved digital signal processing.

Description

[Technical field]

[0001] The present invention relates to a method for improving digital signal processing in a high frequency system, and further to a system as well as a data carrier signal. [Background technology]

[0002] Digital signal processing, or "DSP" for short, can be used in high frequency systems for various applications such as filtering and / or detection of specific conditions of the system. To improve processing and performance, the DSP can be at least partially implemented in hardware, for example by an integrated circuit, or "IC" for short. However, parameterization of an integrated circuit may prove to be time consuming and technically complex. The complexity of the aforementioned parameterization process may affect the digital signal processing in many undesirable ways. One of them is the reduced efficiency of the partial DSP in the integrated circuit.

[0003] In a plasma power supply system, a short RF signal with high frequency can be measured. Typical RF measurements include forward and reverse power, as well as RF current and voltage at various points in the plasma power supply system. The measured signal may generally vary in frequency, amplitude, noise or spurious content and pulse length. Some variations, such as frequency, may be under good control of the power supply system, while other variations may not be controlled, especially when delivering RF power to a plasma processing chamber. The variations may cause various types of interference in the measured RF signal. For example, harmonics as one form of interference may be generated by the nonlinearity of the plasma itself. Another form of interference may result from different frequencies used for the generation of the RF signal. In advanced plasma systems, multiple RF generators with different frequencies may be used for plasma excitation. However, in addition to the desired generation of the intended set of frequencies, mixed products such as additive or subtractive combinations of the fundamental frequency with their harmonics may also be generated.

[0004] Many interferences found in the measured RF signal can be suppressed or at least reduced by using appropriate filters before digitizing the measured RF signal. Nevertheless, undesirable sidebands close to the base frequency may occur, which may require further filtering by the DSP or impose further requirements on filtering. Real input signals are never ideal, as they necessarily contain unavoidable distortions or other artifacts caused by the signal source, the transmission path, or the measurement equipment. Optimization of the filters or filter parameters is important to reconstruct a distortion-free, ideal input signal for stable regulation of the network. However, each filter has advantageous and disadvantageous characteristics. For example, if the signal delay due to filtering becomes too large, the control will not be able to respond in a timely manner to deviations from the intended power level. Therefore, not only the filter type but also its parameter configuration must be carefully selected to balance these characteristics and achieve optimal performance.

[0005] In general, insufficient interference reduction can cause fluctuations in the measurement of the output power of an RF generator. These fluctuations can be understood as "measurement artifacts". Since the measurement can be used for various applications such as the control of the output power level, these applications can depend on a faulty measurement signal. With regard to the control example, the output power level can erroneously follow the measurement artifacts instead of an accurate measurement.

[0006] US20060262889A1 discloses a power supply that can generate a signal when driven by a drive signal from a control device. At high frequencies, either synchronous undersampling or synchronous oversampling may apply.

[0007] In DE102015212242A1 / US20180122625A1, the sampling frequency is changed to remove interference.

[0008] Both publications describe specific methods for compensating for interference frequencies that involve adaptation of the sampling frequency rather than machine learning, in particular the optimization of filter parameters by machine learning procedures.

[0009] Furthermore, it is known from the prior art that intelligent systems exist for detecting certain situations during plasma processing, however, the detection often lacks the reliability and processing speed to initiate corrective action before the completion of the event.

[0010] The use of dedicated solutions for the detection of specific conditions in plasma processing has been reported in the prior art, for example in US8989888B2.

[0011] From the prior art, for example from WO2007095574A1 or DE102011103861A1, the general use of programmable circuits, in particular field programmable gate arrays (FPGAs), for accelerating simulations is known. Summary of the Invention [Problem to be solved by the invention]

[0012] It is therefore an object of the present invention to at least partially overcome the above mentioned drawbacks, in particular to improve digital signal processing in RF systems. [Means for solving the problem]

[0013] The above-mentioned object is solved by a method having the features of claim 1, a system having the features of claim 16 and a data carrier signal having the features of claim 19. Further features and details of the invention are disclosed in the respective dependent claims, the description and the drawings. Features and details explained in connection with the method according to the invention also apply to the system according to the invention as well as to the data carrier signal according to the invention and vice versa in each case, so that cross-reference can always be made with regard to the disclosure relating to the individual aspects of the invention.

[0014] In contrast to prior art publications in the prior art, the method, system and carrier signal of the invention make it possible to optimize digital signal processing, in particular in high frequency systems. For this purpose, it is in particular the idea of ​​the invention that a processing procedure can be used in particular for training the inventive system. Training may be carried out in a training mode with reference data so that an optimized parameterization of the programmable circuits and / or algorithms is determined, and then the inventive system may be switched to an application mode in which the optimized parameterization is used for processing the relevant input signals in real time. This programmable circuit may be a programmable signal processing circuit such as an FPGA for processing the input signals by executing algorithms.

[0015] Further advantages of the method, system and carrier signal of the present invention over the prior art can also be achieved by performing training and real-time operation modes on the same system. Moreover, further training with a new set of training data, i.e. machine learning of new situations, can be advantageously performed at a later time point, and the existing parameterization can be further optimized (hereinafter, this is specifically called "switching the processing from real-time mode to training mode"). In addition, the transfer of the optimized parameterization to other systems of the same kind is possible, for example, by using a user interface.

[0016] Advantageously, the use of the same hardware components (FPGA, control, memory) for both training and application modes facilitates the transfer of parameters between systems, allowing future upgrades of the plasma processing system to new, as yet unknown, situations. Training can be performed, for example, on a test bench system or a laboratory plasma processing system using existing input signals recorded on the system used in the field. The variability of these signals can be numerically increased by scaling, filtering, superposition, etc. The optimized parameters can then be easily transferred from the training system to the plasma processing system in the field.

[0017] Furthermore, by using existing input signals from the memory, the functionality of the RF system in various situations can be tested, for example during factory acceptance testing of the RF system, where the input signals are provided by means of the memory instead of or in combination with the ADC data, and the system behavior, for example the performance of the regulation loops, the filtering of high frequency signals, pattern recognition, as well as any response actions, for example alarms or interlocks, can be verified with a fully operational system.

[0018] This object is in particular solved by a method for improving digital signal processing, preferably in a high frequency system, in particular in a high frequency power supply system, comprising (in particular performed one after the other or in any order): providing at least one input signal, preferably related to a radio frequency signal of the radio frequency system and / or to a signal from the plasma chamber and / or to an optical signal measured in the radio frequency system and / or to a signal of an external source; - executing a processing procedure, in which the at least one input signal is iteratively processed by a circuit, preferably a programmable circuit, using at least one configurable parameter of the processing, the configuration of the at least one parameter being changed, in particular modified, for each processing of the at least one input signal to obtain a respective processing result; Determining at least one parameter result for optimizing the configuration based on the processing result, including in particular the selection of the modified configuration; Providing at least one determined parameter for improved digital signal processing, in particular for parameterization of the circuit using the selection.

[0019] The method according to the invention has the advantage that a number of processing results are available, each specific to a parameter configuration used, so that the best configuration can be selected from the various configurations based on the processing results. In other words, the parameters can be optimized by varying the configuration (also called "varying the parameters"). Thus, the selection of the parameter result can include at least one of the various configurations that gives the best processing result within the processing procedure. By using the parameter result for the parameterization of the circuit, an improvement in digital signal processing (DSP) can be realized.

[0020] The processing steps may include linear iterative processing, in particular digital signal processing, but may also include further steps such as evaluation and / or modification of the configuration, as will be explained in more detail below. Furthermore, the processing procedure may be considered, at least in part, as a machine learning procedure. In other words, the processing procedure may include a machine learning procedure. The machine learning procedure may include (at least) evaluation and / or storing in a memory a determined parameter result, and / or determining and / or providing a parameter result, and / or learning and / or optimizing a configuration, preferably parameter values ​​and / or at least one weight of a neural network.

[0021] In the context of the present disclosure, indefinite articles or numerical designations, such as "one", "two", etc., are to be understood as "at least" designations, unless otherwise stated. Even from "at least" or a similar express designation, it should not be concluded that a limitation in the sense of, for example, "exactly one", should be implied by another simple use of the article or numerical designation without this express designation. Furthermore, numerical designations and designations of process parameters and / or device parameters are to be understood in the technical sense, i.e. with normal tolerances.

[0022] The use of the term "best processing results" is: Optimal conditions for digital signal processing, e.g. maximum suppression of unwanted radio waves, Optimal results for measurement parameters related to the RF power delivery system, e.g. RF power output with least variation; Optimal detection of a particular pattern in one of the measured signals; Optimal results for measured parameters related to plasma processing, e.g. DC self-bias or ion concentration, and It refers to the optimum result of a measured parameter related to a property of an item being processed in a plasma processing chamber, for example, the thickness of a deposited layer.

[0023] The execution of the processing procedure may be initiated and controlled by a control circuit (hereinafter also referred to as "control") electrically connected to the programmable circuit for providing at least one input signal. Furthermore, the determination and providing may be performed by the control circuit.

[0024] The term "digital signal processing" is generally understood to mean the processing of digital signals. In particular, digital signal processing according to the present invention can include the processing of RF signals after they are detected by at least one sensor and then converted in real time into digital signals by an analog-to-digital converter, ADC for short, of the RF system. In contrast to digital signals, the "processing procedure" can be considered as a training procedure that does not use input signals that are actually sensed in real time (instead, the input signals are pre-recorded or simulated by a computer). The algorithms used for processing in the processing procedure and in the digital signal processing can be the same. However, the configuration of the algorithm is first trained in the processing procedure and only applied in the digital signal processing afterwards.

[0025] The term "RF system" refers to a system that outputs at least one electrical RF signal. Furthermore, an RF device may comprise or be part of an RF generator. The term radio frequency power supply system may be a specific configuration of an RF system and refers to a power source for a plasma system, in particular a plasma generator system. A plasma system is suitable for generating plasma used in plasma processing, preferably plasma etching or plasma deposition. As is well known to those skilled in the art, plasma processing applications are generally powered with (amplified) power signals of 100 W or more, often also above 1 kW, to adequately ionize gases and generate plasma in a plasma processing chamber. However, the control signals involved in controlling these high power signals may include much lower power signals and may be processed according to low power signal processing techniques. Sensing the characteristics of a high power signal requires various types of sensors and may involve picking up features of that signal, including a small fraction of the high power signal's amplitude.

[0026] The "at least one input signal" may preferably be a digital signal related to a high-frequency signal of a high-frequency system. In particular, this means that the at least one input signal is characteristic of an RF signal and / or may be obtained by sensing, in particular measuring, an RF signal. The RF signal may therefore be a physical signal in the high-frequency range that is the actual electrical output of or in the RF system. The RF range may be a frequency range from 20 kHz to 300 GHz, preferably from 300 kHz to 300 MHz (medium frequency band, high frequency band and very high frequency band). The input signal may also be a DC signal derived from the RF signal and that may vary at low frequencies, for example a measured DC bias voltage present in a pulsed application, for example a measured forward and / or reflected power, a reflection coefficient, a phase between RF voltage and current, or a scalar signal from an external sensor.

[0027] In particular, the RF range may be in the range of 300 kHz to 500 kHz, or 1 MHz to 3 MHz, or 10 MHz to 16 MHz, or 24 MHz to 30 MHz, or 37 MHz to 43 MHz, or 57 MHz to 63 MHz, or 77 MHz to 83 MHz, or 159 to 165 MHz. For example, in the RF range of 10 MHz to 16 MHz, the specific frequency of the high frequency signal may be 13.56 MHz. The above frequency bands are particularly used for plasma applications and / or RF generators.

[0028] In this processing procedure, at least one input signal is iteratively processed by a circuit. This circuit may be suitable for executing an algorithm for processing the at least one input signal depending on the configuration of at least one parameter. The parameters may therefore be parameters of the algorithm. In case of filtering, the algorithm may be a filter algorithm and the parameters may be filter parameters and / or filter coefficients. Furthermore, the circuit may be suitable for applying a neural network. In this case, the parameters may be at least one weight of the neural network. The circuit may be at least programmable in the sense that the configuration of the at least one parameter may be changed.

[0029] "Iterative processing" may be understood to mean repeatedly processing the same at least one input signal using the same algorithm but with different configurations. Thus, the configuration of at least one parameter is changed for each processing of the at least one input signal to obtain a respective processing result. In other words, each processing result may be a respective result of the algorithm when executed in a respective configuration.

[0030] Advantageously, the at least one parameter may be a processing parameter that influences the processing result. If the processing comprises filtering of the at least one input signal, the at least one parameter may comprise at least one filter parameter. For example, the at least one parameter may comprise an adjustable cut-off frequency and / or a threshold value, etc.

[0031] "Determining at least one parameter result" may be understood as selecting at least one or exactly one of the various configurations that results in the best processing result (and thus is based on the processing result). This allows for optimization of the parameter configurations. "Providing at least one determined parameter for improved digital signal processing" may be understood as parameterization of the circuit using the selected configuration.

[0032] The parameter result may include a particular configuration selected from various configurations, the selection being made depending on the processing result. For example, a criterion such as a metric or a reference result may be provided that is used to evaluate each of the processing results. Each processing result corresponds to a configuration that was used to obtain that processing result. As a result, each of the processing results may be used to evaluate the configurations, especially the speed. Thus, the exact configuration that corresponds to the best evaluation or best evaluation processing result may be selected. That is, the configuration is optimized such that the difference between the corresponding processing result and the criterion is minimized. The "optimization of the configuration" may also be referred to as "optimization of at least one parameter", and the at least one optimized parameter (i.e., the selected configuration) may then be used to parameterize (or another) programmable circuit for "in-situ" or "real-time" operation or "application mode" of the high frequency system. These terms are used synonymously below.

[0033] "Evaluation" may be understood as a comparison of the processing result with a reference, such as an ideal processing result. The better the processing result and the reference match, the better the evaluation of the corresponding configuration. The evaluation may be performed by a control circuit or by a dedicated part of a programmable circuit. There are many different state-of-the-art methods that can be used for the evaluation, for example statistical methods for signal evaluation, multivariate statistical techniques, or empirical comparative time series analysis.

[0034] In one embodiment, the at least one parameter result includes at least two different parameter results for different applications of the DSP. Alternatively or additionally, the DSP and signal processing may include filtering that is parameterized by configuration. The selection of the filtering configuration significantly affects the interference reduction in the RF signal. However, one particular configuration may not work for all combinations of RF power levels in a multi-frequency application, and in some cases may not work for all processes performed in the plasma chamber of the plasma system. Therefore, it may be necessary to determine an optimized configuration for different plasma processing conditions. This can be achieved by using multiple parameter results for different applications. For each application, a different set of input signals is provided during the training process and can be used separately for the processing procedure, resulting in different parameter results.

[0035] Depending on the sequence of different process conditions in a step of a plasma processing recipe, fast switching between different configurations may be advantageous when the process conditions are changed, for example, during real-time operation, and therefore different parameter results may be stored in memory for fast switching.

[0036] Within the digital signal processing and / or processing procedure, at least one input signal may be predefined, for example, by providing an already digitized signal that serves as a training input signal or as a reference input signal. The already digitized signal may be an existing signal, a prerecorded signal, or, for example, a simulated signal. Thus, an ADC is not necessarily required for processing. However, in real-time operation, an ADC may convert an analog sensed RF signal into a digital representation of the signal (hereinafter referred to as a real-time input signal). In training mode, at least one predefined input signal is predefined with desired characteristics and features for the processing procedure, and thus may "represent" only real radio frequency signals. However, in real-time mode, the programmable circuit may process a real-time input signal from a sensed radio frequency signal instead of a predetermined input signal.

[0037] The RF system may comprise at least one sensor for performing detection, in particular measurement, of the RF signal used to supply RF power to the plasma system. The measurement may include measuring the RF signal, in particular in the form of forward and reverse power and / or RF current and / or RF voltage, at various points in the RF system. It is also intended to adjust the output power of the RF generator, and for this purpose the RF signal, in particular in the form of an RF waveform at the output of the generator, is measured by a sensor. A directional coupler may be used to provide the forward and reverse waveforms to the sensor. The RF signal may thus be an analog signal, which is then converted by an ADC with a certain sampling rate to provide a digital real-time input signal. Similarly, other sensors, for example phase and magnitude sensors, voltage-current sensors or DC bias sensors in an impedance matching network or directly in the plasma chamber, for example Langmuir probes or optical sensors, may measure various parameters such as the radio frequency voltage and current, the DC self-bias of the plasma or the intensity of certain spectral lines in the plasma spectrum. These measurements are used to control tunable elements to maximize the power transfer to the plasma or to characterize the plasma properties. The measured parameters may be provided as real-time input signals or may be represented as training input signals.

[0038] Depending on the RF frequencies present in the RF signal, aliasing and / or intermodulation effects may occur when the signal is digitized, i.e., converted by an ADC. These effects may include mixing and / or folding back of high frequency components to lower bands, depending on the sampling rate and frequency. Furthermore, the mixing and / or folding back to lower bands may be changed by the pulse modulation and / or frequency tuning of the RF generator. As a result, the measurement of the delivered RF power at the operating frequency may be affected, and the feedback regulation of the generator that relies on this measurement may deliver inaccurate or unstable or fluctuating power levels. Similarly, it may cause inappropriate reactions such as an unstable output of the RF generator or detuning of the matching network. However, the problem may be mitigated by using processing with programmable circuits, in particular filtering, for example via a suitable low-pass filter. For example, filtering may be used to remove high frequency components above the Nyquist limit (i.e., half the sampling frequency). Unwanted additive or subtractive frequency intermodulation products may be removed, for example, by suppressing certain frequencies or frequency regions after transforming the RF signal into the frequency domain, for example by a fast Fourier transform. Thereby, the parameter result may include at least one (optimized) configuration for at least one parameter, and thus in the case of filtering, this configuration may be called a filter configuration.

[0039] The type and layout of filters required for the most accurate measurement results after digitization and / or mixing, as well as the filter coefficients, may vary for amplitude modulated signals, in particular depending on the specific time characteristics of the multi-level pulse patterns, e.g., pulse rise and fall times, power difference between levels, pulse frequency, etc. One set of filter coefficients that works for a particular group of pulse patterns may not be appropriate for significantly faster or slower pulse patterns. By using filter configurations optimized individually for each pulse pattern, unwanted aliasing and / or intermodulation products can be suppressed for various pulse patterns. In the case of frequency tuning, for example, to adapt to a variable plasma load and to maximize the power delivered to the plasma system, variable fundamental frequency dependent filter coefficients may be required to obtain accurate measurements and to maintain a stable output power of the RF generator. By implementing the filter components and optimized parameters in programmable circuits, i.e. in particular programmable logic devices, e.g., digital signal processors (DSPs), complex programmable logic devices (CPLDs), or FPGAs, aliasing and / or intermodulation products can be removed or reduced in real time. Removing aliasing and / or intermodulation products is only possible to a certain extent inside programmable circuits, in particular FPGAs. Moreover, the implemented filter configuration can be quickly switched to another configuration optimized for different operating conditions. Moreover, if at a later point in time an improved filter configuration is developed, it can be easily implemented without changing the hardware. In this context, it is already known that aliasing and / or intermodulation effects are bypassed by sampling the signal in such a way that undersampling effects are avoided by adapting the sampling frequency to the signal base frequency. In contrast, the inventive use of a specific configuration for processing, in particular filtering, allows to remove higher frequencies, while the base frequency can still be sampled at a sufficiently high rate.Moreover, implementation in a programmable circuit allows for fast switching between different configurations, for example, depending on a particular application or situation. An optimized parameter set for a given application case can be obtained, for example, by teaching the system with at least one training input signal, which can include a reference waveform, and their expected reference results. Then, the configurable parameters, for example, filter parameters, can be varied until an optimal value is reached. Note that the anti-aliasing filter, which is placed in front of the ADC, is an analog circuit. This anti-aliasing filter can be made variable using controllable analog components, but cannot be implemented in an FPGA.

[0040] The at least one input signal may include a plurality of predefined or pre-existing input signal groups (i.e., pre-recorded or, for example, simulated signal groups). Each of the input signals may be related to a radio frequency (RF) signal, i.e., may be specific to or representative of an actual or simulated RF signal. They may also be easily derived from a measured RF signal, for example, by generating a digital representation from the measured RF signal. Furthermore, each input signal may include a known pattern, and the corresponding reference result may indicate this known pattern, as will be described in more detail below. Since the pre-existing input signals may include certain characteristics and / or provide corresponding reference results, they may be considered as reference groups or reference signals. Optimization of the configuration, i.e., learning the optimal parameterization of the programmable circuit, may be achieved by comparing the reference results with the process results. The reference results may include target waveforms or target frequencies or desired pattern detection results. The comparison may be achieved by using an appropriate metric that indicates the degree of discrepancy between the reference results and the process results.

[0041] After providing at least one input signal in the form of a group of input signals, i.e. with (at least) a first input signal and a second input signal, in particular with at least 10 or 100 or 1000 further input signals, the processing procedure, which may also be called a training procedure, may comprise the following processing flow: a first processing step comprising processing a first input signal using the first configuration, processing a second input signal using the first configuration, and processing a further input signal using the first configuration to obtain a first processing result for the first configuration; a second processing step comprising processing a first input signal using a second configuration, processing a second input signal using the second configuration, and processing a further input signal using the second configuration to obtain a second processing result for the first configuration; In a further processing step, the first input signal is processed using the further configuration, the second input signal is processed using the further configuration, and the further input signal is processed using the further configuration, thereby obtaining a corresponding processing result for each of the further configurations.

[0042] Thus, the configuration of each run can be changed by repeating the run for different parameter configurations. For training procedures, the change in configuration (and therefore the change in parameters) can be done by established methods, e.g., gradient descent or backpropagation.

[0043] It is advantageous if the programmable circuit is configured as a programmable integrated circuit, preferably a digital signal processor (DSP), complex programmable logic device (CPLD) or field programmable gate array (FPGA). Alternatively, it may be configured as a fixed circuit, such as an application specific integrated circuit (ASIC), or other integrated circuit. In general, any computer or controller (e.g. PC, microcontroller, DSP, CLPD, FPGA, etc.) can be used for machine learning, in particular by machine learning procedures. It is particularly suitable for fast real-time digital signals, which can speed up the machine learning procedures. Furthermore, the use of programmable circuits, in particular FPGAs, allows flexible and automated training. Many applications of high frequency systems, such as plasma applications, evolve in different ways, and FPGA solutions have the advantage that future applications requiring adaptive processing, for example filter designs that are not yet known, can still be handled with the same hardware architecture. In contrast, fixed circuits, for example ASICs, need to be modified or completely redesigned. This increases the amount of work and the costs. The method according to the invention allows for easy addition of new configurations, preferably via software upgrades that may include more advanced filter configurations. The user can then select one of these configurations for a new application.

[0044] Furthermore, it is optionally possible within the scope of the present invention that the method further comprises: using the at least one determined parameter result to parameterize the same or another programmable circuit; and Detects high frequency signals from high frequency systems, performing said digital signal processing using processing of said detected radio frequency signals sensed at least in part by said parameterized circuit; Monitoring the high frequency signal by repeatedly performing sensing and digital signal processing, preferably the processing procedure can be called a training procedure used to perform parameterization of the programmable circuit in a training mode of processing using an existing signal as a training input signal, and / or the processing procedure can be retrieved from memory and configured with a parameterization that may have been previously determined in the training mode, while in particular the monitoring is performed in a real-time processing mode, in which the high frequency signal is sensed and processed as a real-time input signal.

[0045] This means that the high frequency system can actually be used practically in real time and can be switched from training mode to real time mode.

[0046] At a later point in time, the radio frequency system may be switched back from real-time mode to training mode and the parameterization may be further optimized by iterative processing of additional existing signals, for example previously recorded or simulated signals.

[0047] In particular, it is an advantage of the present invention that the programmable circuitry allows both real-time processing of an input signal and training of the device by iteratively processing a pre-existing training signal to determine the optimal configuration for real-time processing. The pre-existing signal may be a pre-recorded signal or, for example, a computer-simulated signal.

[0048] These two features make the RF system very versatile and can be trained to handle new plasma processing conditions or detect new situations without requiring hardware and algorithmic changes; only the parameterization of the programmable circuitry is changed. Instead of having to use an external computer to analyze measured signals and determine improved configurations, such optimized configurations can be determined by the actual RF system itself by switching the actual RF system into training mode. In this way, the programmable circuitry of the RF system has the properties of a machine learning capable system, or of an on-device inference framework, both of which can be adapted to new applications.

[0049] It is a further advantage of the present invention that the programmable circuits used, in particular FPGAs, are much faster, e.g. 10,000-100,000 times faster, in performing digital signal processing compared to standard computers and standard model simulations using tools like Matlab / Simulink.

[0050] The parameter result may include an optimized parameter set, i.e. each configuration of at least one parameter optimized using the processing procedure, so that the programmable circuit can be parameterized using this parameter set of the parameter result. As a result, the method according to the invention achieves improved digital signal processing by using a highly accelerated optimization of at least one parameter in a first step. In a second step of real-time operation, the processing procedure and the programmable circuit are used, which results in a faster and smarter selection of the parameter set for the digital signal processing. For digital signal processing of the sensed high-frequency signal during real-time operation, the programmable circuit can receive at least one real-time input signal converted by an analog-to-digital converter from the sensed high-frequency signal.

[0051] Furthermore, the circuitry for at least the parameter optimization can be the same as the circuitry for the final application in real-time operation of the high frequency system.

[0052] Furthermore, within the scope of the present invention, at least one input signal processed in the processing procedure is configured as at least one training input signal, the method further comprising the steps of: determining a parameterization of the programmable circuit by configuring at least one parameter using the predetermined parameter results, the configuration preferably being subsequently fixed; switching the processing from the training mode to a real-time mode, in particular in the real-time mode, the real-time input signal is processed by the programmable circuit in real time using a fixed configuration, in particular for filtering the real-time input signal, preferably configured as an actual measured signal received from at least one analog-to-digital converter which converts a high-frequency signal sensed by at least one sensor into at least one real-time input signal, Switching processing from real-time mode to a training mode, preferably in the training mode, further optimizing the existing previously fixed configuration by running the processing procedure again using at least one additional training input signal as a processed training input signal to newly determine at least one parameter result, wherein re-parameterization of the programmable circuit is performed by configuring at least one parameter using the newly determined parameter result in response to a comparison of the newly determined parameter result with the previously determined parameter result.

[0053] This allows further optimization of the existing parameter set. The user can decide whether to switch back to the training mode to load more reference input signals and reference results of the existing configuration. The loading or uploading can be performed via an electronic interface or a user interface. Alternatively or in addition, additional reference input signals and reference results may be recorded by the RF system itself and directly loaded into the storage device in real-time operation to be used later in the training mode. The old and new sets of parameter configurations are then evaluated, and the better parameter set is then stored in memory. Furthermore, switching from the real-time mode to the training mode allows training of a new situation at a later time, i.e., switching to the training mode, loading new reference input signals and reference results into memory, determining an optimized parameter set, and storing it in memory.

[0054] The user can also instruct the RF system via the electronic interface when the results of the training with the new reference input data are to be used to optimize the existing parameter configuration, i.e., to overwrite the existing configuration with the optimized configuration, or when the optimized parameter configuration is to be used for a new application, and in particular to be stored as a separate parameter configuration in memory to be used for the new application.

[0055] According to another embodiment of the invention, the iterative processing of the at least one input signal is performed by iterative processing steps, where for each step the at least one input signal is processed by the same programmable circuit while using a modified configuration of at least one parameter, so that a respective processing result is specific and / or assigned to the modified configuration, in particular the at least one input signal is provided as one or more predefined reference input signals, for each of which a reference result representative of a desired result of the processing of the corresponding input signal is provided, and determining the at least one parameter result for the optimization of the configuration on the basis of the processing results comprises: Evaluating the changed configuration by comparing each of the processing results obtained for at least one reference input signal with a corresponding reference result, preferably where the evaluation depends on matching the processing result with the reference result and in particular includes identifying deviations between the processing result and the reference result. In other words, the parameters can be optimized by changing the configuration. Iterative processing of the same at least one input signal is used, which can apply different configurations in the sense of different parameter values. This leads to different processing results, possibly slightly. Each of the processing results can be used to evaluate the performance of the particular configuration used to obtain said processing result.

[0056] In particular, the reference groups, reference input signals and / or reference results mentioned in connection with the present invention are also referred to as reference data.

[0057] Furthermore, the use of the evaluation can provide an optimization of the configuration in the sense that the parameter result can include a parameter set that results in a processing result that deviates least from the reference result for all reference input signals. The optimization of the configuration can also be called parameter optimization. The processing procedure can therefore be called a training and / or optimization procedure. At least one predefined reference input signal and the corresponding reference result can be referred to as training data. The determined at least one parameter result can then be verified by performing the processing and the evaluation of the processing result again for a second set of reference data (including the new input signal and the corresponding reference result). In this way, the problem of overfitting the parameter configuration to the training data is avoided.

[0058] Within the scope of the present invention, determining at least one parameter result for optimizing the configuration based on the processing result may include: selecting at least one configuration having a highest rating as the at least one determined parameter result; In particular, providing at least one determined parameter for improved digital signal processing includes: parameterizing the same or another programmable circuit using the at least one selected configuration; and Preferably, the at least one predetermined reference input signal and corresponding reference result comprises a plurality of predetermined reference input signals and corresponding reference results for different applications of digital signal processing, such that determining the at least one parameter result for optimizing a configuration based on a processing result is performed separately for each application, and preferably providing the at least one determined parameter result comprises storing a selected configuration in a memory, and parameterizing the same or another programmable circuit comprises selecting a configuration to use from the stored configurations depending on the application.

[0059] In other words, different training data sets can be provided for different digital signal processing applications, which allows switching between parameter sets for different applications.

[0060] A further possibility is that the at least one input signal is configured as a multiplicity of pre-existing input signals (a "signal group"), preferably the processing procedure comprising iterative processing steps, each processing step comprising: The same input signal is processed one after the other, a control circuit for configuring at least one parameter used in this step; The method includes: evaluating, for each of the processing steps, a deviation of the processing result from a predetermined reference result by the control circuit; after each evaluation, modifying a configuration of at least one parameter used for a subsequent processing step based on said evaluation; Preferably, the control circuit is electrically connected to the programmable circuit and configured separately from the programmable circuit; In particular, the configuration of at least one parameter used is changed for different processing steps to iteratively optimize the configuration to improve the digital signal processing.

[0061] The control circuit or control unit may be configured as an additional control processor separate from the programmable circuit, with a connection to a memory containing training data, e.g. at least one input signal and / or a number of existing input signals and / or corresponding reference results. The existing input signals may be pre-recorded signals or, e.g., simulated signals. The control circuit may also be implemented on the programmable circuit. The programmable circuit may be an FPGA. Providing the at least one input signal, in particular the selection of training data for a processing procedure (e.g., based on the application to be trained) and / or the modification of the configuration within the processing procedure may not be performed directly by the programmable circuit itself, but by the control circuit. For example, at each iteration of the processing, the control circuit may provide the input signal and / or the modified configuration to the programmable circuit. The control circuit may then perform an evaluation and thus determine the next modification of the configuration depending on the evaluation. It is also possible that the evaluation is at least partially performed by the programmable circuit in order to accelerate the training procedure.

[0062] Alternatively or additionally, the programmable circuit can have a direct connection to the shared memory, or at least a part of the memory can be implemented by the programmable circuit. The control circuit can then instruct the programmable circuit from which address range of the memory to load the training data. The programmable circuit can then directly access the memory, load and process the training data one after another, and evaluate the processing results. This method has the advantage of significantly accelerating reading from the memory.

[0063] In the context of the present invention, it may further be provided that the processing procedure is configured as a training procedure for optimizing the configuration in an iterative manner for different specific situations in the application of the radio frequency system, so that the parameter results determined for the different situations are configured as different optimized parameter sets for these different situations. The different situations may be different applications of digital signal processing. The different parameter sets may include different configurations adapted to these different situations. This allows flexible switching between the parameter sets depending on the specific situation. Thus, the training procedure allows training of different situations, resulting in optimized parameter sets depending on the different situations.

[0064] In another preferred embodiment, at least one input signal comprises a measurement of a time signal path, preferably instantaneous and historical values, in particular of the high frequency signal or a signal derived therefrom, and a processing result is obtained by processing the signal path and indicates the possibility of detecting a certain pattern in the signal path. Thus, not only the instantaneous values ​​of the input signal are taken into account, but also a series of values ​​from previous time intervals can be processed and the probability of detection, i.e. the expected value, can be calculated. In training mode, the instantaneous and historical values ​​can be predefined and / or prerecorded. However, in real-time mode, the real-time input signal can also comprise a measurement of a time signal path, preferably instantaneous and historical values, in particular of the high frequency signal or a signal derived therefrom. That is, in real-time mode, not only the instantaneous measurements of the high frequency signal are taken into account, but also a series of values ​​from a predefined time interval of the previous measurement point. This allows the use of the processing in different applications, such as probabilistic evaluations.

[0065] At least one input signal, i.e. the training input signal and / or the real-time input signal, may be configured as at least one digital signal. The digital signal may be provided by an analog-to-digital converter that converts the sensed analog RF signal into a digital signal. The digital signal may be a digitized representation of the sensed RF signal and may therefore be treated as a one-dimensional array of data points describing the instantaneous amplitude of the measured RF signal. In case of continuous measurements over time, the array may be further expanded such that the most recent data points represent the most recent measurements. The temporal measurement rate of the measurements, and therefore the temporal measurement rate of the expansion of the array, may depend on the sampling rate of the digital-to-analog converter. By choosing a time window of a certain length that always ends with the most recent data point, it is possible to obtain a sequence of one-dimensional arrays that are updated at the measurement or sampling rate of the digital-to-analog converter by shifting out old data points from one end and adding new data points to the other end of the array.

[0066] In particular, real-time measurements are also required to identify critical situations and address them before they develop into real undesirable phenomena. Such methods are particularly useful for handling overvoltage or overcurrent conditions in electronic equipment or arcing phenomena during plasma processing. As a result, continuous analysis of a stream of measurement data for the presence of certain patterns may be required.

[0067] Advantageously, the method according to the invention further comprises the detection of at least one pattern in the sensed high frequency signal and / or at least one input signal. To improve the reliability of the detection of a particular situation, the high frequency signals measured by multiple sensors can be used in combination such that at least one pattern is present in each signal at a particular time. For example, successful detection of a phenomenon may require that a particular pattern is detected in two high frequency signals at the same time or with a particular time delay between them. False positive detections due to measurement artifacts or interferences can thus be avoided.

[0068] In particular, "performing digital signal processing by at least partly using processing of a sensed high-frequency signal by a parameterized circuit" and / or "performing a processing procedure in which at least one input signal is repeatedly processed by a programmable circuit using at least one configurable parameter of the processing" respectively include the detection of at least one pattern in the sensed high-frequency signal and / or in the at least one input signal. The sensed high-frequency signal may be configured as a real-time input signal and therefore as a digital signal representation as described above. The at least one input signal is therefore also used within the processing procedure and may be configured as a training input signal and therefore also as a digital signal representation as described above. In both cases, the processing may be designed to provide for the detection of at least one pattern, preferably in the form of pattern recognition. This configuration may therefore be a configuration for pattern detection, in particular a configuration of a pattern matching or pattern recognition algorithm and / or the weights of the individual neurons of the neural network. The pattern matching or pattern recognition algorithm has the advantage of being technically less complex than a neural network. However, a neural network can in several circumstances detect patterns more reliably in noisy or fluctuating signals than simple pattern matching. A combination of both is also possible. The configurable parameters may therefore be configured as parameters of algorithms or neurons of a neural network. The processing procedure may also be configured as or comprise or be part of a machine learning procedure for machine learning and / or training of a neural network, and the digital signal processing may relate to the application of the trained neural network. The at least one parameter result for the optimization of the configuration may further be configured as a training result in the form of a set of neuron weights. It may be provided that the occurrence of the phenomenon is detected and / or predicted based on the detection of at least one pattern.

[0069] Furthermore, it may be provided that an action is initiated based on the detection or prediction. In particular, the result of the pattern detection is a probability of occurrence of the pattern. Additionally or alternatively, the detection and / or prediction based on the detection includes a determination of a probability of occurrence of the phenomenon. The initiation of the action may depend on a comparison of this probability with a threshold value. This makes it possible in particular to start the operation before the end of the phenomenon. The action may be a corrective action to counteract the occurrence of the phenomenon, for example a shutdown of the RF system.

[0070] In more advanced embodiments, the probabilistic detection or prediction of the phenomenon can be based on multiple conditions. This can be done by monitoring the progression of the probability of the occurrence of the phenomenon over a period of time. Rather than simply detecting the phenomenon by comparing the instantaneous probability value to a threshold, additional conditions for more reliable detection can be, for example: A probability value that is consistently above a specified threshold for a specified period of time; The average probability value of exceeding a specified threshold for a specified period; increasing the probability value after the probability value exceeds a specified threshold; The derivative of the probability value within a specified period of time of exceeding a threshold; Correlation of probability values ​​determined for several monitoring signals; Any combination of one or more of the above and further conditions.

[0071] With the help of more complex conditions, the detection of the phenomenon can be more reliable and less dependent on the momentary fluctuations of the probability. Advantageously, the monitoring period of the probability value can be selected short enough so that the phenomenon is still suppressed before its end by initiating the action. Thus, the training of the system can include a parameterization for the detection of the phenomenon as early as possible and / or for the classification of the phenomenon as accurately as possible. This result can then receive a higher weighting in the parameterization.

[0072] A particular pattern may be detected in at least one digital signal, particularly in the form of a line array as described above. Preferably, the detection of the pattern is trained using a training input signal as the digital signal and / or is performed in real time using a real time input signal as the digital signal. This allows for identifying a phenomenon during operation of the RF system based on the detection of the pattern. The phenomenon may be related to a trouble occurring in the RF system or the plasma processing system, requiring an action such as shutting down the RF system. Furthermore, the phenomenon may indicate a degradation of a particular component of the RF system or the plasma processing system. The phenomenon may be notified to a user via a particular analog or digital signal, which may be provided, for example, via an electronic or user interface, so that the user can respond to the situation with additional corrective measures or schedule repairs and / or preventive maintenance at a later time.

[0073] The particular patterns may include interferences and / or irregularities and / or characteristic changes as described above. These patterns may be related to particular phenomena in plasma processing, plasma processing systems, or RF power supply systems. If the RF signals are repeatedly detected by successive measurements, the digital signals represent the successively detected signals. By continuously processing the digital signals, particularly the resulting sequence of one-dimensional arrays, the particular patterns can be monitored for changes over time.

[0074] Furthermore, by using an artificial neural network, a fully evolved detection of a particular pattern can be learned by optimizing the weights of the individual neurons, and with the optimized set of weights, the network can recognize the occurrence of a particular pattern in real time. An appropriate output function of the neural network can be selected that generates a probability of a particular pattern in a real-time input signal or data stream in order to detect an event before it has fully ended.

[0075] The same result is achieved by calculating the cross-correlation between a reference pattern and a stream of actual measured signal data, the reference pattern being successively shifted in time relative to the signal waveform.

[0076] If the phenomenon is fully unfolded, the pattern is recognized with a probability close to 1. However, before that point, the calculated probability is gradually increasing from 0 towards 1. As soon as the probability exceeds a certain user-adjustable threshold, or a combination of several conditions as mentioned above, the phenomenon can be reported before its actual termination. In this way, irregularities in the stream of measurement data that correlate with undesirable or even disastrous phenomena can be flagged before the occurrence of the phenomenon. In this case, measures can be initiated to avoid the phenomenon's full unfolding and to reduce and / or suppress serious consequences. The probability threshold and further conditions as mentioned above can be set separately for each type of phenomenon. For example, catastrophic phenomena can be treated with a low threshold, while small irregularities can be treated with a higher threshold. The advantage of using a lower threshold for serious irregularities is that there is more time to carry out corrective measures before the phenomenon is fully unfolded. On the other hand, a higher threshold allows the phenomenon to be classified with a higher accuracy.

[0077] For example, arcing on the surface of a semiconductor wafer in a plasma processing system can cause serious defects and loss of one or more chips on the wafer. If the arc is strong enough, the plasma processing sequence can be strongly disturbed, which can result in the entire wafer being scrapped. Therefore, for such catastrophic arcing events, a low probability threshold is beneficial, so that the event can be suppressed early. In another example, a high probability threshold may be more appropriate for less severe events. In plasma processing chambers used for deposition processing, not only the semiconductor wafer but also parts of the plasma chamber are coated with a deposition layer. After processing many wafers, these chamber parts are covered with an increasingly thick layer of deposition material into which particles can fall. During plasma processing, such particles can cause plasma irregularities, but can be tolerated to a certain extent. Nevertheless, it is important to accurately classify and count all particle-related events so that preventive maintenance, in this case, cleaning of the plasma chamber, can be initiated if the number of particle-related events exceeds a threshold. In this case, a high probability threshold allows particle-related events to be separated from other irregularities, so that premature cleaning of the chamber caused by erroneous particle event counting can be avoided.

[0078] It may be easy to implement the optimized neural network into a programmable logic device, such as an FPGA, since all parts of the neural network can be transferred to the corresponding logic components. Similarly, pattern matching algorithms can be easily implemented in an FPGA by using shift registers and simple logic operations. Changing the configuration and / or size of the neural network or modifying the pattern matching algorithm can be done by loading a new configuration into the FPGA without changing the hardware.

[0079] If an application requires the detection of multiple pattern classes, multiple parallel neural networks or pattern matching algorithms can be set up, each optimized for a specific class. It is also possible to use one common neural network or pattern matching algorithm to detect multiple pattern classes. In this case, instead of one probability value as the output of the network, an array of probabilities is generated, one for each class. Further pattern class detection can be added later. If necessary, the parameterization of the neural network or pattern matching algorithm can be optimized by additional teachings.

[0080] This allows different configurations to be stored for different applications and activated depending on the situation. In particular, fast switching between different configurations is possible, for example activated one after the other for different successive process steps. Subsequently, activating a teaching mode on the installed system may allow further optimization of existing configurations or learning new situations with additional reference data. Thus, probabilistic pattern recognition in the measured signal is also possible by applying parallel (simultaneous) classification to multiple pattern types, even before the patterns are fully deployed.

[0081] Furthermore, an electronic interface may be provided for uploading at least one input signal or an existing configuration to the device of the invention, in particular to the programmable circuit, or to the control unit, or to the memory. The method of the invention then uses the uploaded existing configuration to further parameterize the programmable circuit. If the input signal is uploaded to the control unit, the control unit may load the signal into the memory before providing it to the programmable circuit. The at least one input signal may be configured as a plurality of input signals that are stored in the memory after uploading.

[0082] The electronic interface may be configured as a user interface allowing the user to upload at least one input signal and / or a user input signal and / or a user configuration and / or a control signal allows the system to switch from training mode to real-time mode in order to use a particular configuration stored in the memory for parameterization of the programmable circuit. The uploaded input signals are generally formulated and the training data may be provided to the programmable circuit directly or via the control unit, or they may be stored in the memory, directly or via the control unit. The user input signal refers to at least one input signal from a user. The user configuration may be an existing configuration that the user loads into the interface, for example via cloud or on-site storage. The user configuration may be a configuration created by using the method of the invention on the user input signal. In general, the training data may include at least one input signal and a corresponding reference result. The training data may be configured as existing (previously recorded or simulated) data, especially for a training procedure, which may be initially loaded into the memory and processed one by one (in various iterations) until an optimized parameter set is found. The uploaded existing configuration can be stored in the same memory and / or in a different memory and / or the uploaded existing configuration can overwrite a configuration already implemented according to any of the methods described above. When the uploaded existing configuration is stored in the same memory, it can act as a supplementary / additional parameterization of the programmable circuit. The advantage of the uploaded configuration being stored in a different memory or memory space is that a backup configuration is always at hand. It can also be advantageous for the manufacturer to be able to limit access to any configuration not executed by the user. In this case, there is also the option to load any of the configurations into a safe and trusted memory space.The advantage of overwriting the configuration of a programmable circuit by uploading another configuration is the new parameterization of the programmable circuit and the reduction in memory space required.

[0083] An electronic interface may be provided to access the parameter configuration and / or the fixed configuration and / or at least one input signal and / or data from the memory and / or the programmable circuit and / or the control unit. However, it is also possible that the access via the electronic interface is restricted, for example so that reading and / or overwriting and / or replacement of certain data and / or algorithms or neural networks is prohibited. Thus, part of the data or information stored in the memory and / or the programmable circuit and / or the control unit can be locked, for example by using encryption methods. In other words, it is possible to use the electronic interface to perform an upload of configurations to be stored in the memory for use in a specific application. These configurations may be the result of a training procedure performed on a similar circuit with appropriate training data. In particular, the manufacturer may allow the uploading of configurations to the user, but at the same time the basic factory configuration, as well as the algorithm or neural network itself, may be protected from access by the user. This has the advantage that the user can still revert to the factory configuration if he is not satisfied or if the system is no longer working with the new configuration.

[0084] By using only the programmable circuit, the user interface, the memory and the control as a training bench, it is possible to perform training and determination of the optimal parameter set for the parameterization of the circuit. As in the training mode of the complete radio frequency system, the already digitized training data is read from the memory and provided by the control as input signals for digital signal processing. The evaluation of the change of parameters and the results of the processing is also performed in the same way as for the complete RF system. This training bench system with a common commercial FPGA can determine the optimized parameter results in a much shorter time, for example 10,000-100,000 times faster, compared to a standard model simulation using a standard computer and tools like Matlab / Simulink.

[0085] The optimal parameterization of common filtering algorithms used in high frequency systems can be determined within minutes using the training bench system described above, compared to run times that can take hours or days when using Matlab / Simulink simulations.

[0086] Even for large neural networks with thousands of neurons arranged in, for example, 500-1000 layers, the training bench system can determine the optimal set of neuron parameters within a few hours. For even more complex algorithms or neural networks and / or for very large training input data sets, the determination of the optimal configuration of the programmable circuitry can be performed on a dedicated high-performance supercomputer.

[0087] Independently of the above possibilities of determining the optimal configuration on the RF system itself, on a training bench system, on a standard computer using simulation tools, or on a high-performance supercomputer, the determined configuration can be uploaded by the control into the memory of the RF system via a user interface and used to parameterize a programmable circuit for a specific application, which can then process the input signals of the RF system in real time.

[0088] Similarly, a configuration determined in one RF system may be exported via a user interface and uploaded to a similar RF system used to parameterize the programmable circuit. In this way, by implementing an existing configuration from a first RF system on a second system of the same type, an individual training of the second system can be avoided. Thus, a copy of an already optimized configuration can be used on several systems, which has the advantage of saving time when setting up a new system.

[0089] Although unavoidable small variations between systems may require further optimization of the parameterization of the new RF system, this optimization of parameters should nevertheless typically converge faster starting from a configuration already known to be good compared to starting from a default parameter configuration, which may be the case for an entirely new application.

[0090] At least one input signal may be configured as a predetermined reference input signal, for each of which a desired result of the processing is provided as a corresponding predetermined reference result. The training data may also include the predetermined reference input signal and the corresponding predetermined reference result.

[0091] The circuit may execute an algorithm parameterized by at least one configurable parameter of the circuit to provide processing of at least one input signal. The interface, in particular the user interface, may be used for selection of an appropriate configuration for the parameters before the process is started. Alternatively or additionally, the interface may be used for switching from an application or real-time operating mode to a training mode. The processing procedure may be configured as a learning cycle for optimization of the configuration. The interface may also be controlled by the user to load reference input signals and / or reference results and / or training data into a memory, in particular the memory of the high frequency generator, and / or to perform a learning cycle and / or to store a new parameter configuration in the memory for later use in the application mode. In the training mode, the memory may be used to simulate a continuous data stream from the analog-to-digital converter from the application mode. The memory may thus be connected to a control circuit, which may read at least one input signal from the memory and provide it to the programmable circuit in the training mode. The programmable circuit may also read at least one input signal directly from the memory to perform signal processing in order to accelerate processing. On the other hand, in an application or real-time operating mode, the programmable circuit can be directly connected to an analog-to-digital converter to receive the input signal as a real-time input.

[0092] It is a further option that the at least one input signal is configured as at least 1000 or at least 1 million different input signals, where each of the input signals represents at least one waveform related to the high frequency signal, where the processing is aimed at evaluating the waveforms in response to the parameter configuration, and each processing result includes a result of said evaluation. This can further improve the performance of the processing. Furthermore, automatic training, for example controlled by the control circuit, can accelerate the optimization of the at least one parameter.

[0093] In another advantageous embodiment, the programmable circuit comprises logic elements for decomposing the input signal into amplitude and phase using an algorithm, the algorithm being parameterized by various configurations. Thus, an algorithm such as IQ demodulation or the Goertzel algorithm or the discrete Fourier transform depends on many parameters. For example, in the case of IQ demodulation, the parameters can be the coefficients of a digital filter. The configuration of parameters is therefore the specific values ​​selected for these parameters. Further details can be found in the description of FIG. 3 below.

[0094] Furthermore, the evaluation of the various configurations is performed by comparing each of the processing results associated with the various configurations with a predetermined reference result, which may comprise a predetermined ideal result of the algorithm executed by the programmable circuit. The predetermined ideal result is therefore a measure of the quality of the currently set configuration. A change in the parameter configuration makes it possible to generate different processing results that can be compared with the reference result. The deviation between the processing result and the reference result indicates the performance of the used configuration, which may be understood as the evaluation of the configuration.

[0095] Another aspect of the invention is a system, in particular a radio frequency measurement system for a radio frequency system, preferably a radio frequency power supply system or an RF generator or a radio frequency plasma system, comprising: at least one sensor for sensing a high frequency signal of the high frequency system, in particular a directional coupler or a voltage-current ("V / I") sensor; at least one analog-to-digital converter for converting the at least one detected high frequency signal into at least one input signal; A circuit, in particular a programmable circuit, for digital signal processing of at least one input signal, said circuit being parameterized by at least one parameter result determined according to the method according to the invention.

[0096] The system according to the invention therefore provides the same advantages as those described in detail with reference to the method according to the invention. The device may also comprise an analog anti-aliasing filter and / or a matching circuit. The radio frequency system may in particular be configured as a radio frequency power supply system. The radio frequency power supply system may further be configured as a power supply for the plasma system, in particular for the plasma power supply system. An analog-to-digital converter may be used for the power supply and may be provided for converting the sensed radio frequency signal sensed by the sensor of the radio frequency system.

[0097] The device may further comprise an electronic interface for uploading an existing configuration to the programmable circuit, or to the control for parameterization of the circuit, or to the memory. Uploading and / or writing and / or reading using the interface may be performed using data transfer. The electronic interface may be electronically connected to the control or the memory or the programmable circuit to perform the upload using (in particular direct) data transfer. The electronic interface may be configured as a data interface, such as a serial interface, which may be connected to another electronic device, such as a computer. It is also possible for the electronic interface to execute a computer program for enabling the data transfer. The computer program may provide for more sophisticated interface technologies, such as a local area network, a wireless local area network, or Bluetooth. The electronic interface is accessible via the control or the memory or the programmable circuit.

[0098] The system according to the invention can be characterised by any of the above-mentioned features of the method according to the invention and further exhibits the advantageous features mentioned above.

[0099] Another aspect of the invention is a data carrier signal carrying at least one parameter result determined according to the method according to the invention. The data carrier signal according to the invention thus offers the same advantages as the corresponding features described in detail for the method and system according to the invention. The data carrier signal can be a signal on a data carrier such as a hard disk or a flash memory. Furthermore, the data carrier signal can be implemented by downloading etc. A further possible aspect of the invention is a computer program comprising the parameter result determined according to the method according to the invention and / or a computer program comprising the parameter result determined according to the method according to the invention and / or a computer readable data carrier stored on the computer program. The data carrier signal may be provided for performing a software upgrade on the user side, i.e. for parameterizing the programmable circuit using the determined parameter result of the data carrier signal. Alternatively, the data carrier signal may be provided for performing a parameterization / reading of the programmable circuit, so that the configuration may be stored in the memory of the RF system or exported, i.e. stored, on a removable data carrier medium connected to the user interface of the RF system. The removable data carrier medium may then be connected to a second RF system or a training bench system in order to upload the exported configuration and use it for parameterization of the programmable circuitry of the second RF system.

[0100] Further advantages, features and details of the invention will become apparent from the following description in which an embodiment of the invention is explained in detail with reference to the drawings, in which: In this connection, the features mentioned in the claims and in the description may each be essential to the invention individually or in any combination. [Brief description of the drawings]

[0101] [Figure 1a]1 shows an embodiment of an RF measurement system 1 according to the invention and a data carrier signal according to the invention. [Figure 1b] 1 illustrates an embodiment of a plasma power system that uses the RF measurement system 1 of the present invention. [Diagram 2] 1 illustrates one embodiment of the method of the present invention for improving DSP in a high frequency system of the present invention. [Diagram 3] The operating principle of the programmable circuit of the present invention used in a DSP is shown. [Figure 4] 1 illustrates an embodiment of an RF measurement system of the present invention in a training mode configuration. [Diagram 5] 1 illustrates an embodiment of an RF measurement system of the present invention in an application (real-time operation) mode. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0102] In the following drawings, the same reference numbers are used for the same technical features even in different embodiments.

[0103] FIG. 1a shows an embodiment of the inventive RF measurement system 1 of a high-frequency generator, preferably a high-frequency power supply system 1, or a high-frequency generator or a high-frequency plasma system. The system comprises at least one sensor 20, in particular a directional coupler or a voltage-current sensor or a plasma sensor 32, for sensing the high-frequency signal of the high-frequency system 1. The directional coupler 20 is, for example, a type of sensor commonly used to probe high-frequency signals in power supply circuits. However, other sensors can also be implemented. FIG. 1a further shows a high-frequency RF power signal transmission line 60 and two branches, the upper branch can represent the forward power and the lower branch representing the reflected power of the RF power signal picked up by the sensor 20. The structure of the inventive system 1 described herein can be suitable for either the measured forward power or the reflected power, or both. The structure of the inventive system 1 described herein can also be suitable for the measured RF voltage, RF current, reflection coefficient (Γ or "gamma value"), or the measured DC bias, or any combination of RF phase and AC current (measured forward power or reflected power, or both) and DC bias and reflection coefficient.

[0104] The system 1 further comprises at least one analog-to-digital converter and an ADC 30 for converting at least one sensed high-frequency signal into at least one input signal 230. The ADC converts a time-continuous and value-continuous, aka analog signal, into a time-discrete and value-discrete, aka digital signal. The analog signal may be the one picked up by the directional coupler 20. The number of samples per second is given by the sampling rate and the number of different values ​​is given by the number of bits of the ADC 30. The system 1 may further comprise an analog anti-aliasing filter 50 and a matching circuit 40 between the sensor and the ADC 30.

[0105] The system 1 comprises a circuit 10, preferably programmable, for digital signal processing. The circuit 10 is suitable for digital signal processing of at least one input signal 230. The input signal 230 can be a measurement signal during operation of the RF generator, i.e. a real-time mode signal 230. The circuit is further suitable for digital signal processing of at least one input signal 210 stored in a memory 70. Either the input signal 230 or the stored input signal 210 can be a signal related to a high-frequency signal of the power supply system. They can therefore be directly measured or simulated signals. The stored input signal 210 can be further processed by a control or control circuit 80 before entering the circuit 10. The circuit 10 can be further parameterized by at least one parameter result determined according to the method of the invention described in FIG. 2. The parameterization can be performed by the time any input signal 210 or 230 reaches the circuit 10. The parameterized circuit 10 then outputs a processing result 220 according to the input signal (the input signal from the memory 210 or the real-time mode 230). The processed results 220 may then be passed through the control circuitry 80 and stored in the memory 70 as parameter results 240 .

[0106] FIG. 1a also shows a data carrier signal 240 carrying at least one parameter result 240 determined from the process result 220. The data carrier signal 240 can be a data package that can be downloaded by an authorized person or any third party. The data carrier signal can then be used to parameterize the programmable circuit 10, in particular the FPGA 10. The process result 220 used for the parameter result 240 can be determined from the method of the invention, according to at least one of the methods described herein. The system can further comprise an electronic interface 90 for uploading an existing configuration to the programmable circuit 10, or to the control unit 80, or to the memory 70 for parameterization of the circuit 10. Said electronic interface 90 can be electronically connected to the control unit 80, the memory 70, or the programmable circuit 10, to perform the upload using direct data transfer. Furthermore, the electronic interface 90 can also be used to perform a download, for example to transfer the configuration to a different RF system for a similar application. The electronic interface 90 can be a user interface.

[0107] 1b illustrates an embodiment of a plasma power system 5 that uses the inventive RF measurement system 1. In particular, the plasma power system 5 includes an RF generator 2, a matching network 4, and a plasma processing system 3, e.g., a plasma chamber 31 having an electrode through which RF power is supplied.

[0108] The RF generator 2 comprises an RF measurement system 1 comprising a sensor 20, an analog-to-digital converter 30, a programmable circuit 10, a control unit 80, a memory 70 and an electronic interface 90, also called a user interface 90. Furthermore, the RF generator 2 comprises a signal generating unit 16, e.g. a direct digital synthesis (DDS) circuit and a power amplifier stage 23, powered by a DC power supply 21, the signal generating unit 16 being able to be implemented within the programmable circuit 10. It should be noted that within the scope of the present invention, the signal generating unit 16 may also be a separate circuit or chip provided outside the programmable circuit 10.

[0109] Matching network 4 includes an input sensor 41, e.g., a phase and magnitude detector, an impedance matching circuit 40, typically including inductors and capacitors that may be variable, a match control 42, and, optionally, an output sensor 43, e.g., a voltage-current sensor or a DC bias sensor. The function of matching network 4 is to adapt the output impedance of RF generator 2, typically 50 ohms, to the variable impedance of plasma processing system 3, such that the RF power coupled into the plasma is maximized and the power reflected back from plasma processing system 3 is minimized.

[0110] The plasma processing system 3 comprises a plasma chamber 31 to which RF power is supplied from a matching network 4. In the plasma chamber 31, a gas or gas mixture is excited and partially ionized by the supplied RF power. Electrical coupling of the RF power to the plasma can be achieved by connecting the RF power to an electrode, for example a metal plate opposite a similar plate connected to ground. In this way, the plasma is capacitively coupled to the two electrodes. Alternatively, the RF power can be connected to an inductor arranged around the plasma chamber 31 so that the plasma is inductively coupled.

[0111] Typically, a semiconductor wafer is loaded into the plasma chamber 31 and exposed to an RF plasma for a specified period of time to etch material from or deposit material onto a top layer of the wafer.

[0112] The matching network 4 and the plasma processing system 3 may include sensors 41, 43, 32 that measure RF signals or signals characteristic of properties of the plasma processing. Such sensor signals may be connected to dedicated sensor ports of the RF generator 2 and converted to a digital representation by an analog-to-digital converter (ADC) 30, e.g., a single-channel and / or multi-channel ADC, and used as input signals for digital signal processing performed by the programmable circuit 10.

[0113] 2 shows an embodiment of the method of the invention for improving the DSP in a high-frequency system of the invention, in particular in a high-frequency power supply system 1. The method comprises providing 101 at least one input signal 210 related to a high-frequency signal of the high-frequency system 1 and executing 102 a processing procedure 310, where the at least one input signal 210 is repeatedly processed by the programmable circuit 10 using at least one configurable parameter of the processing. The configuration of the at least one parameter thereby can be changed or specifically modified for each processing of the at least one input signal 210 to obtain a respective processing result 220. The method further comprises determining 103 at least one parameter result for the optimization of the configuration (e.g., specifically an optimized parameter from the varying parameters) based on the processing result 220 and providing 104 at least one determined parameter result for improved digital signal processing (specifically for line parameterization). According to the invention, the programmable circuit 10 can be configured as a programmable integrated circuit 10, preferably as an FPG.

[0114] The invention may further comprise parameterization 105 of the same or another programmable circuit 10 using at least one determined parameter, the parameterization 105 comprising sensing 106 a radio frequency signal of the radio frequency system 1 and performing 107 a digital signal processing by using processing of the sensed radio frequency signal by at least a part of the parameterized circuit 10. The invention may further comprise monitoring the radio frequency signal by repeatedly performing 102 the sensing 106 and the digital signal processing. Regardless of the details of the parameterization 105 of the circuit 10, the processing procedure 310 of the programmable circuit 10 may be configured as a training procedure for performing the parameterization 105 of the programmable circuit 10 in a training mode of processing using an input signal 210 as an existing (e.g. previously recorded or simulated) training input signal 210, while the monitoring is performed in a real-time mode of processing in which a radio frequency signal is sensed and processed as a real-time input signal 230.

[0115] The at least one input signal 210 processed in the processing procedure 310 may be configured as at least one training input signal 210. The method of the present invention may further include (1) finishing the parameterization 105 of the programmable circuit 10 by configuring at least one parameter using a previously determined parameter result, where the configuration is subsequently fixed, and (2) switching of processing from a training mode to a real-time mode, where the real-time input signal 230 is processed in real-time by the programmable circuit 10 using the fixed configuration, in particular for filtering of the real-time input signal 230, where the real-time input signal is subjected to at least one analog-to-digital conversion for converting a high-frequency signal sensed by the at least one sensor 20 into the at least one real-time input signal. and (3) switching of processing from a real-time mode to a training mode, in which the existing and previously fixed configuration can be further optimized by running the processing procedure 310 again to newly determine at least one parameter result using at least one additional training input signal 210 as the processed training input signal 210, where in response to a comparison of the newly determined parameter result with the previously determined parameter result, a re-parameterization of the programmable circuit 10 is performed by configuring at least one parameter using the newly determined parameter result.

[0116] The iterative processing of the at least one input signal 210 may be performed by iterative processing steps, where for each step the at least one input signal 210 is processed by the same programmable circuit 10 using a modified configuration of at least one parameter, so that the respective processing result 220 is specific to and / or assigned to the modified configuration. The at least one input signal 210 may be configured as one or more predefined reference input signals 257, for each of which a reference result is provided that represents a desired result of the corresponding input signal 210. The determination 103 of the at least one parameter result for the optimization of the configuration based on the processing result 220 may further comprise an evaluation of the modified configuration by comparing each of the obtained processing results 220 for the at least one reference input signal 257 with the corresponding reference result, the evaluation depending on a match between the processing result 220 and the reference result or a deviation between the processing result 220 and the reference result. The determination 103 of the at least one parameter result for the optimization of the configuration based on the processing result 220 may further comprise selecting the at least one configuration having the highest evaluation as the at least one determined parameter result. The reference result can be provided by the user, for example uploaded using the user interface 90. Thus, the user can define at any time which criteria are used for the evaluation. Also, the input signal, and thus the training data, may be defined by the user and uploaded, for example, via the user interface 90 for that reason. Providing 104 at least one determined parameter for improved digital signal processing may further comprise parameterizing 105 of the same or another programmable circuit 10 using at least one selected configuration. Preferably, said at least one predefined reference input signal 257 and corresponding reference result comprises a plurality of predefined reference input signals 257 and corresponding reference results for various applications of digital signal processing, so that determining said at least one parameter result for optimizing a configuration based on the processing result 220 is performed separately for each application.Providing 104 the at least one determined parameter result may further include storing the selected configuration in a memory, and parameterizing the same or another programmable circuit 10 includes selecting a configuration to use from the stored configurations depending on the application.

[0117] The at least one input signal 210 can be configured as a number of existing input signals 210. The existing signals can be pre-recorded or can be simulated, for example. The processing procedure 310 can include iterative processing steps, each processing step including successively processing the same input signal 210 and setting, by the control circuit 80, the configuration of at least one parameter used in this processing step. The invention can further include, after each processing step, evaluating, by the control circuit 80, the deviation of the processing result 220 from a predefined reference result, and, after each evaluation, modifying, based on said evaluation, the configuration of at least one parameter used for the next processing step. Thereby, the control circuit 80 can be electrically connected to the programmable circuit 10 and configured separately from the programmable circuit.

[0118] The configuration of the at least one parameter used can be varied for different processing steps in order to iteratively optimize the configuration for improved digital signal processing.

[0119] The processing procedure 310 can be configured as a training procedure for optimizing the configuration in an iterative manner for different specific situations in the application of the high frequency system 1, and thus the parameter results determined for the different situations are configured as different optimized parameter sets for these different situations.

[0120] At least one input signal 210 may comprise a time signal path, preferably from instantaneous and historical values, in particular from measurements of high frequency signals. A processing result 220 is obtained by processing the signal path and may indicate the probability of detecting a particular pattern in the signal path.

[0121] An electronic interface 90 may further be provided for performing uploading of the at least one input signal 210 to the programmable circuit 10, whereby the at least one input signal 210 may be configured as multiple input signals 210 that are stored in the memory after uploading.

[0122] The at least one input signal 210 can be configured as at least 1000 or at least 1 million different input signals 210, where each of the input signals represents at least one waveform related to a high frequency signal. The processing here can aim at evaluating the waveforms according to the parameter configuration. Each processing result 220 can include the result of said evaluation.

[0123] An existing configuration can also be uploaded via the electronic interface 90, in particular into the programmable circuit 10, or into the control unit 80, or into the memory 70. The parameterization of the programmable circuit 10 then uses the existing configuration. The uploaded existing configuration can be stored in the same memory 80 and / or in a different memory. The uploaded existing configuration can overwrite any of the configurations calculated using the method of the invention, or can be stored in the memory in addition to any other configurations calculated using the method of the invention or previously uploaded and stored in the memory.

[0124] The programmable circuit 10 may comprise logic elements for decomposing the input signal 210 into in-phase and quadrature components (I / Q demodulation) or amplitude and phase using algorithms, which may be parameterized by various configurations.

[0125] The evaluation of the modified configuration may be performed by comparing each of the processing results 220 associated with the modified configuration to predetermined reference results. The predetermined reference results may include predetermined ideal results of the algorithm executed by the programmable circuit 10.

[0126] FIG. 3 shows a data flow graph of an IQ demodulator that can be implemented in an FPGA as a programmable circuit 10 according to the invention. Logic elements are provided that allow performing mathematical operations on the digital input signals 210, 230. The digital signals 210, 230 can be decomposed into amplitude and phase by arranging the logic elements. The decomposition can be performed via different algorithms such as IQ demodulation, the Goertzel algorithm, or the Discrete Fourier Transform (DFT). The example shown in FIG. 3 shows the decomposition of the digital input signals 210, 230 using IQ demodulation, which results in the in-phase part of the signal being available at point I and the out-of-phase (or quadrature) signal being available at point Q. The basic operations can be the squaring 14 and summing 15 of the digital signals to reconstruct the amplitude of the original signals 210, 230 and provide it at point A. Here, the sum of the squared value of the in-phase signal and the squared value of the out-of-phase signal is equal to the squared value of the amplitude. The FPGA 10 further comprises a local oscillator 11 that provides an internally generated digital signal at a given frequency and phase. For example, the local oscillator 11 can provide a value of f(t)=sin(ω1t). Since the FPGA is programmable, any frequency ω1 between 0 and half the sample frequency can be generated by the local oscillator 11 according to the Nyquist theorem. The frequency can also be swept from one value to another during successive time intervals. Then, for I demodulation, the FPGA performs a multiplication 12a of the internally generated digital signal from the local oscillator 11 with the input digital signal 210 or 230. For Q demodulation, the FPGA performs a multiplication 12b of the digital input signal 210 or 230 with the internally generated digital signal from the local oscillator 11 with an offset φ=90°. This offset ensures that if the multiplication 12a for I demodulation is performed with a sine function, the multiplication 12b for Q demodulation is performed with the corresponding cosine function. The products obtained by the multiplications 12a, 12b are then processed in a digital filter 13. The filter can have adjustable parameters, for example to adjust the slope of the filter acting beyond the cutoff frequency. Any of the decomposition algorithms depends on several parameters.In the case of IQ demodulation, the coefficients of the digital filters are as described above. The operations performed on the digital signal to extract information can reveal the mean over time, the standard deviation, or various other statistics, depending on the desired outcome of the digital signal processing. Furthermore, it can be simulated on a computer using a general-purpose programming language such as Python, or specialized tools such as the Xilinx add-on system generator for Matlab-Simulink. The computation time on a standard computer exceeds the simulation time by several orders of magnitude. Thus, the above-mentioned inventive method using programmable circuits, especially FPGAs, can significantly reduce the simulation run time.

[0127] 4 illustrates an embodiment of the RF measurement system 1 of the present invention configured to operate in a training mode, for example as a training bench system. The operating mode 253 of the system 1 can be set to the training mode via the user interface 90.

[0128] In the following, the term "signal processing circuitry" designated by the reference number 17 refers to that part of the programmable circuit 10 dedicated to performing signal processing. The programmable circuit 10 may comprise further circuits, for example an evaluation circuit 18 used in the training mode for determining the deviation between the processed result 220 and a reference result. The further circuits used during the real-time mode may be implemented in the programmable circuit 10 but may be deactivated in the training mode.

[0129] When in training mode, the power amplification stage 23, the impedance matching network 4, and the connection of the system 1 to the plasma processing system 3 are not mandatory. Instead of real-time input signals provided by the RF generator 2, the matching network 4, or sensors in the plasma processing system 3, the system 1 can process existing reference input signals 257, which are stored in the memory 70 or uploaded via the user interface 90. Furthermore, reference results 256 can be stored in the memory 70 or provided via the user interface 90. These reference results indicate expected results of signal processing of the reference input signals 257. Also, parameter configurations 254 that can be used to parameterize the signal processing circuit 17 can also be stored in the memory 70 or provided via the user interface 90.

[0130] Additionally, algorithms for digital signal processing, e.g., neural networks, or circuit layout 255, that may be implemented in programmable circuit 10 may also be stored in memory 70. Typically, uploading of algorithms and circuit layout 255 via user interface 90 and their implementation in programmable circuit 10 is restricted to the manufacturer and is not accessible to normal users of system 1.

[0131] In the training mode, the controller 80 can be configured to read the user parameter set 251, the user reference results 252a, and the user reference input signals 252b from the user interface 90 and store them in the memory 70. Additionally, the controller 80 can be configured to parameterize the programmable circuit 10 by reading the stored parameter set 254 from the memory 70 and parameterizing the programmable circuit 10 accordingly.

[0132] Furthermore, the control unit 80 can be configured to read the reference input signal 257 from the memory 70 or the user interface 90 and provide it to the signal processing circuit 17 as the input signal 210. Alternatively, the signal processing circuit 17 may directly access the memory 70 and directly read the parameter configuration 254 and the training data, i.e., the reference input signal 257 and the reference result 256, so that the reading and processing of the training data is accelerated.

[0133] 4 may be implemented within the programmable circuit 10 or may be implemented as a software algorithm within the control unit 80. The result of the evaluation, i.e., the optimized parameter set 254, may be stored in the memory 70 by taking it from the programmable circuit 10 and writing it to the memory 70 via the control unit 80 or by direct memory access from the programmable circuit 10.

[0134] 4, it should be noted that at least a part of the control unit 80 and / or at least a part of the memory 70 can be implemented in the programmable circuit 10. This has the advantage that the reference input signal 257 and the reference result 256 can be read from the memory 70 and the subsequent signal processing of the reference input signal 257 and the evaluation of the processing result 220 against the reference result 256 can be significantly accelerated.

[0135] 5 shows an embodiment of the inventive RF measurement system 1 in application (real-time operation) mode. In application mode, the signal processing circuit 17 is parameterized with an existing parameter set 254 stored in memory 70 during a previous training or uploaded to memory 70 via user interface 90. Depending on a process recipe 258 that may be provided via user interface 90, for each recipe step, a dedicated parameter configuration may be loaded into the signal processing circuit 17 from memory 70 or from user interface 90 via control 80. The signal processing circuit 17 thus parameterized processes in real time input signals continuously provided from the RF system, e.g., from the RF generator 2, the impedance matching network 4, or sensors in the plasma processing system 3, and digitized in the ADC 30.

[0136] Depending on the algorithm or neural network implemented in the programmable circuit 10, the result of the signal processing can be used to control a signal generator 16, e.g., a DDS core, or further circuitry 19. The signal generator 16 is used, for example, to adjust the amplitude, frequency, and phase, etc. of a radio frequency signal that is converted into an RF power signal by a digital-to-analog converter (DAC) 22 and a power amplifier stage 23. The resulting RF power signal at the output of the RF generator 2 can then be provided to the plasma processing system 3 via an impedance matching network 4. In this way, for example, the result of filtering the input signal 210 from the directional coupler 20 at the output of the RF generator 2 can be directly used to adapt the RF power signal, for example, so that the RF power delivered to the plasma is stabilized and the reflected power from the plasma is minimized.

[0137] In addition, the signal processing circuitry 17 may also be configured to monitor the input signal 210 for patterns indicative of certain phenomena or undesirable conditions, such as high voltage discharges in the plasma processing system 3, or degradation of hardware components related to plasma processing, such as particle buildup in the plasma system, or drift in plasma impedance indicating that cleaning or other preventative maintenance of the plasma chamber 31 may be required. Such phenomena may be flagged (259) to a user via the user interface 90, or may be acted upon directly by further circuitry 19 in the programmable circuitry 10, which may trigger corrective action 250. For example, in the case of an arcing event or high voltage discharge, or if it is determined that such an event is likely to occur, the programmable circuitry 10 may initiate a reduction or complete turning off of RF power for a period of time. This may be achieved, for example, by reducing the amplitude of the RF signal generated in the signal generating section 16, or by disabling the DAC 22, or by interrupting the connection between the DAC 22 and the power amplifier stage 23 for a period of time.

[0138] The circuit layout of the programmable circuit 10 or the algorithms used by the programmable circuit 10 may be stored in memory 70 and different circuit layouts 255 may be implemented in the programmable circuit 10 depending on the application.

[0139] As described above for the training mode, also in the application mode at least part of the control unit 80 and / or part of the memory 70 may be implemented within the programmable circuit 10. This has the advantage that switching between different configurations of the signal processing circuit 17 can be significantly accelerated.

[0140] The above description of the embodiments describes the present invention only based on examples. Of course, the features of each embodiment can be freely combined with each other as long as it is technically reasonable and does not deviate from the spirit of the present invention. [Explanation of symbols]

[0141] 1: system, 2: RF generator, 3: plasma processing device, 4: matching network, 5: plasma power system, 10: circuit, 11: local oscillator, 12: multiplication, 13: digital filter, 14: element operation squaring, 15: element operation addition, 16: signal generation unit, 17: signal processing circuit, 18: evaluation circuit, 19: addition circuit, 20: sensor, 21: DC power supply, 22: digital-to-analog converter (DAC), 23: power amplifier stage (PA), 30: analog-to-digital converter (ADC), 31: plasma chamber, 32: plasma sensor, 40: impedance matching circuit, 41: input sensor, 42: matching control, 43: output sensor, V / I sensor, 50: anti-aliasing filter, 60: power line, 70: memory, 80: control circuit, 90: electronic interface, user interface, 101: providing at least one input signal, first method step, 10 2: execution of second method step, 103: determination of third method step, 104: provision of at least one determined parameter result, fourth method step, 105: parameterization, fifth method step, 106: sensing, sixth method step, 107: execution, seventh method step, 210: (stored / existing) input signal, 220: processing result, 230: (real-time) input signal, 240: parameter result, 250: corrective action, 251: user parameter set, 252a: user reference result, 252b: user reference input signal, 253: operating mode, 254: parameter set, 255: circuit layout, 256: reference result, 257: recorded measurement data set, reference input signal, 258: process recipe, 259: phenomenon flag, 310: processing procedure, A: amplitude, I: in-phase part of signal, Q: out-of-phase part of signal, φ: phase offset

Claims

1. A method for improving digital signal processing in a radio frequency system (1), in particular in a radio frequency power supply system (1), comprising: Providing (101) at least one input signal (210), preferably related to a radio frequency signal of a radio frequency system (1); executing (310) a processing procedure (102), wherein the at least one input signal (210) is iteratively processed by a programmable circuit (10) using at least one configurable parameter of the process, the configuration of the at least one parameter being changed for each processing of the at least one input signal (210) to obtain a respective processing result (220); determining (103) at least one parameter result (240) for optimization of the configuration based on the processing result (220); and providing (104) the at least one determined parameter result for improved digital signal processing.

2. 2. The method of claim 1, wherein the programmable circuit (10) is configured as a programmable integrated circuit (10), preferably a digital signal processor (DSP), a complex programmable logic device (CPLD), or a field programmable gate array (FPGA).

3. Parameterizing (105) the same or another programmable circuit (10) using at least one determined parameter result; Sensing (106) a radio frequency signal of the radio frequency system (1); performing digital signal processing (107) at least in part using the processing of the sensed high frequency signal by a parameterized programmable circuit (10); monitoring the radio frequency signal by repeatedly performing sensing (106) and digital signal processing (102); 2. The method of claim 1, characterized in that the processing procedure (310) is used as a training procedure for performing parameterization (105) of the programmable circuit (10) in a training mode of processing using the input signal (210) as an existing training input signal (210), while monitoring is performed in a real-time mode of processing in which a high-frequency signal is sensed and processed as a real-time input signal (230).

4. The at least one input signal (210) processed in the processing procedure (310) is configured as at least one training input signal (210), the method further comprising: determining a parameterization (105) of the programmable circuit (10) by configuring at least one parameter using the predetermined parameter results, the configuration being subsequently fixed; switching the processing from a training mode to a real-time mode, in which a real-time input signal is processed by the programmable circuit (10) in real time using a fixed configuration, in particular for filtering the real-time input signal, preferably configured as an actual measured signal received from at least one analog-to-digital converter (30) which converts a high-frequency signal sensed by at least one sensor (20) into at least one real-time input signal; 2. The method of claim 1, comprising: switching processing from a real-time mode to a training mode, in which an existing, previously fixed configuration is further optimized by executing the processing procedure (310) again using at least one additional training input signal (210) as a training input signal (210) to newly determine at least one parameter result, and in response to a comparison of the newly determined parameter result with a previously determined parameter result, a reparameterization of the programmable circuit (10) is performed by configuring at least one parameter using the newly determined parameter result.

5. the iterative processing of the at least one input signal (210) is performed by iterative processing steps, and for each processing step the at least one input signal (210) is processed by the same programmable circuit (10) using a changed configuration of the at least one parameter, so that each processing result (220) is specific to and / or assigned to the changed configuration; At least one input signal (210) is provided as one or more predefined reference input signals (257), for each of which a reference result representative of a desired result of processing the corresponding input signal (210) is provided, and determining (103) at least one parameter result for optimization of said configuration based on a processing result (220) comprises:

2. The method of claim 1, comprising: evaluating the changed configuration by comparing each of the processing results (220) obtained for at least one reference input signal (257) with a corresponding reference result, said evaluation depending on a matching of the processing results (220) with the reference result.

6. Determining (103) the at least one parameter result for optimizing the configuration based on the processing result (220) comprises: selecting at least one configuration having a highest rating as the at least one determined parameter result; Furthermore, providing (104) at least one determined parameter for improved digital signal processing may include:

6. The method of claim 5, further comprising: parameterizing (105) the same or another programmable circuit (10) using at least one selected configuration, preferably wherein the at least one predetermined reference input signal (257) and corresponding reference result comprises a plurality of predetermined reference input signals (257) and corresponding reference results for different applications of digital signal processing, such that determining the at least one parameter result for optimizing a configuration based on the processing result (220) is performed separately for each application, and providing (104) the at least one determined parameter result comprises storing the selected configuration in a memory (70), and parameterizing the same or another programmable circuit (10) comprises selecting a configuration to use from the stored configurations depending on the application.

7. The at least one input signal (210) is configured as a number of existing input signals (210), and the processing procedure (310) comprises iterative processing steps, each processing step comprising: processing the same input signal (210) one after the other; setting, by a control circuit (80), a configuration of at least one parameter used in said processing step; The method further comprises: evaluating, after each processing step, the deviation of the processing result (220) by the control circuit (80) from a predetermined reference result; 2. The method of claim 1, further comprising the step of modifying, after each evaluation, a configuration of at least one parameter used for a next processing step based on said evaluation, wherein the control circuit (80) is electrically connected to the programmable circuit (10) and configured separately therefrom, and the configuration of the at least one parameter used is varied for different processing steps to iteratively optimize the configuration for improved digital signal processing.

8. 2. The method according to claim 1, characterized in that the processing procedure (310) is configured as a training procedure for optimizing the configuration in an iterative manner for different specific situations in the application of the radio frequency system (1), so that the parameter results determined for the different specific situations are configured as different optimized parameter sets for these different specific situations.

9. 2. The method according to claim 1, characterized in that the at least one input signal (210) comprises measurements of a time signal path, preferably instantaneous and historical values, in particular high frequency signals, and the processing result is obtained by processing of the time signal path and indicates a possibility of detection of a particular pattern in the time signal path.

10. 2. The method of claim 1, wherein an electronic interface is provided to perform uploading of the at least one input signal (210) to the programmable circuit (10), the at least one input signal (210) being configured as a plurality of input signals (210) that are stored in a memory (70) after uploading.

11. 2. The method of claim 1, wherein the at least one input signal (210) is configured as at least 1000 or at least 1 million different input signals (210), each of the input signals representing at least one waveform related to the high frequency signal, and the processing aims at evaluating the waveforms depending on the parameter configuration, whereby the respective processing results (220) comprise the results of the evaluation.

12. 2. The method of claim 1, wherein the programmable circuit (10) comprises logic elements for decomposing the input signal (210) into amplitude and phase using an algorithm, the algorithm being parameterized by the modified configuration.

13. 2. The method of claim 1, wherein the evaluation of the modified configuration is performed by comparing each of the processing results (220) associated with the modified configuration with predetermined reference results, the predetermined reference results comprising predetermined ideal results of an algorithm executed by the programmable circuit (10).

14. - uploading an existing configuration, in particular to the programmable circuit (10) or the control circuit (80) or the memory (70), via an electronic interface (90); and parameterizing (105) the programmable circuit (10) using the existing configuration.

15. 15. The method of claim 14, characterized in that the uploaded existing configuration is stored in the same memory (70) and / or in a different memory and / or the uploaded existing configuration overwrites any of the configurations according to claim 1.

16. A RF measurement system for a system, in particular a radio frequency system, preferably a radio frequency power supply system (1) or a radio frequency generator or a radio frequency plasma system, comprising: at least one sensor (20), in particular a directional coupler or a current-voltage sensor or a plasma sensor, for sensing a high-frequency signal of the high-frequency system (1); at least one analog-to-digital converter (30) for converting the at least one sensed high frequency signal into at least one input signal (210, 230); a circuit (10) for digital signal processing of said at least one input signal (210, 230), A system, characterized in that the circuit (10) is parameterized by at least one parameter result determined according to the method of claim 1.

17. 17. The system according to claim 16, characterized in that it further comprises an electronic interface (90) for uploading an existing configuration to the circuit (10) or to a control unit (80) or to a memory (70) for parameterizing the circuit (10).

18. 18. The system of claim 17, wherein the electronic interface (90) is electronically connected to the control unit (80) or the memory (70) or the circuit (10) for uploading using direct data transfer.

19. A data carrier signal (240) carrying at least one parameter result (240) determined according to the method of any one of claims 1 to 15.

Images