METHOD FOR DETERMINING A PRESSURE IN A PRESSURE MEASUREMENT CELL AND A MEASUREMENT CELL ASSEMBLY

Abstract

The invention relates to a method and to a measurement cell assembly for determining a pressure in a pressure cell (2) are given, wherein the method consists in that a measurement signal (x) is determined, which is at least proportional to a measured pressure in the pressure cell (2), and in that the measurement signal (x) is filtered by means of a first filter unit (10) having a low-pass characteristic in order to produce an output signal (y), wherein the low-pass characteristics of the first filter unit (10) is defined by means of a first damping factor (α.sub.1). The method is characterized in that an input difference (x_diff), which results from a difference between the output signal (y) and the measurement signal (x), is filtered by means of a second filter unit (20) having a low-pass characteristic to determine an output difference, wherein the low-pass characteristic of the second filter unit (20) is defined by means of a second damping factor (α.sub.2), and in that the first damping factor (α.sub.1) of the first filter unit (10) is determined on the basis of the output difference of the second filter unit (20).

Claims

1. Method for determining a pressure in a pressure cell (2), wherein the method consists of the following: that a measurement signal (x) is determined which is at least proportional to a measured pressure in the pressure cell (2), and that the measurement signal (x) is filtered with a first filter unit (10) having a low-pass characteristic for generating an output signal (y), wherein the low-pass characteristic of the first filter unit (10) is defined by a first damping factor (α.sub.1), characterized in that an input difference (x_diff) resulting from a difference between the output signal (y) and the measurement signal (x) is filtered with a second filter unit (20) having a low-pass characteristic for determining an output difference (y_diff), wherein the low-pass characteristic of the second filter unit (20) is defined by a second damping factor (α.sub.2), and in that the first damping factor (α.sub.1) of the first filter unit (10) is determined on the basis of the output difference (y_diff) of the second filter unit (20).

2. Method according to claim 1, characterized in that the first filter unit (10) comprises a first exponential moving average filter.

3. Method according to claim 1, characterized in that the second filter unit (10) comprises a second exponential moving average filter.

4. Method according to claim 1, characterized in that the measurement signal (x) is present as a discrete-time sampled measurement signal (x.sub.k) at a time represented by a time index (k) and the output signal (y) is likewise present as a discrete-time sampled output signal (y.sub.k) at a time likewise represented by the time index (k), and in that the first and second filter units (10, 20) are implemented as discrete-time filters.

5. Method according to claim 1, characterized in that the low-pass characteristic of the first filter unit (10) and/or the second filter unit (20) is or are first order.

6. Method according to claim 1, characterized in that an absolute value of the output difference (y_diff.sub.k) is determined in an absolute value unit (30), in that the absolute value of the output difference (y_diff.sub.k) is multiplied by a filter effect factor (FW) for determining a product, and in that the product of this multiplication is used for determining the first damping factor (α.sub.1).

7. Method according to claim 6, characterized in that the product of the multiplication is limited to a range from 0 to 1.

8. Method according to claim 6, characterized in that the absolute value of the output difference (y_diff.sub.k) is exponentiated by an exponent (exp) before multiplication by the filter effect factor (FW) is performed.

9. Method according to claim 6, characterized in that the filter effect factor (FW) lies in a value range from 0 to 10.sup.40.

10. Method according to claim 8, characterized in that the exponent (exp) is in a range of values from 0 to 10, typically in a range of values from 0.5 to 5, even more typically equal to 2.5.

11. Method according to claim 1, characterized in that the second damping factor (α.sub.2) is in a range of values from 0 to 1, typically in a range of values from 0.05 to 0.25, even more typically equal to 0.1.

12. Measurement cell assembly having a pressure cell (2) and a membrane pressure measurement cell (2) which is operatively connected to the pressure cell (2) and generates a pressure-dependent measurement signal (x) which is applied to a first filter unit (10) having a low-pass characteristic in order to generate an output signal (y), wherein the low-pass characteristic of the first filter unit (10) is defined by a first damping factor (α.sub.1), characterized in that an addition unit (11) is provided, to which the inverted input signal (x) and the output signal (y) are supplied for determining an input difference (x_diff), in that the input difference (x_diff) of a second filter unit (20) with low-pass characteristic is applied to determine an output difference (y_diff), wherein the low-pass characteristic of the second filter unit (20) is defined by a second damping factor (α.sub.2), and in that the output difference (y_diff) and the second damping factor (α.sub.2) of the second filter unit (20) are applied to generate the first damping factor (α.sub.1) which is supplied to the first filter unit (10).

13. Measurement cell assembly according to claim 12, characterized in that the first filter unit (10) comprises a first exponential moving average filter.

14. Measurement cell assembly according to claim 12, characterized in that the second filter unit (10) comprises a second exponential moving average filter.

15. Measurement cell assembly according to claim 12, characterized in that the measurement signal (x) is present as a discrete-time sampled measurement signal (x.sub.k) at a time represented by a time index (k) and the output signal (y) is likewise present as a discrete-time sampled output signal (y.sub.k) at a time likewise represented by the time index (k), and in that the first and second filter units (10, 20) are implemented as discrete-time filters.

16. Measurement cell assembly according to claim 12, characterized in that the low-pass characteristic of the first filter unit (10) and/or the second filter unit (20) is or are first order.

17. Measurement cell assembly according to claim 12, characterized in that an absolute value unit (30) for determining an absolute value of the output difference (y_diff.sub.k) is present, and in that the absolute value of the output difference (y_diff.sub.k) and a filter effect factor (FW) of a multiplication unit (32) for determining the first damping factor (α.sub.1) are applied.

18. Measurement cell assembly according to claim 17, characterized in that the first damping factor (α.sub.1) in a limiting unit (33) can be limited to a range of values from 0 to 1.

19. Measurement cell assembly according to claim 17, characterized in that a functional unit (31) is provided which is supplied with the absolute value of the output difference (y_diff.sub.k) and with an exponent (exp) for generating a potentized output signal (z.sub.k).

20. Measurement cell assembly according to claim 17, characterized in that the filter effect factor (FW) lies in a value range from 0 to 10.sup.40.

21. Measurement cell assembly according to claim 17, characterized in that the exponent (exp) is in a range of values from 0 to 10, typically in a range of values from 0.5 to 5, even more typically equal to 2.5.

22. Measurement cell assembly according to claim 12, characterized in that the second damping factor (α.sub.2) is in a range of values from 0 to 1, typically in a range of values from 0.05 to 0.25, even more typically equal to 0.1.

Description

[0042] In the following, exemplary embodiments of the present invention are explained in detail with reference to figures, wherein:

[0043] FIGS. 1a and 1b show a measurement cell assembly with a membrane pressure measurement cell connected to a process chamber, with which a measurement signal is determined which, after processing in a signal processing unit according to the invention, is supplied to a valve;

[0044] FIG. 2 shows a block diagram of a first embodiment variant according to the present invention, in particular for implementation in the signal processing unit according to FIG. 1,

[0045] FIG. 3 shows a block diagram of a per se known exponential moving average filter of the first-order low-pass filter type as a discrete-time transfer function, in particular for use in a first and/or in a second filter unit according to FIG. 2, and

[0046] FIG. 4 shows a block diagram of a further embodiment variant of the present invention, in particular for implementation in the signal processing unit according to FIG. 1.

[0047] FIG. 1a shows in a highly simplified block diagram a measurement cell assembly with a process chamber 1, a membrane pressure measurement cell 2, a vacuum pump 3, a signal processing unit 4, a control unit 5, a valve actuator 6 and a valve 7. The membrane pressure measurement cell 2 is used to determine the pressure in the process chamber 1, in which a pressure specified in accordance with a vacuum process is set. Vacuum processes include a wide variety of processes, such as coating processes, etching processes, thermal treatment of workpieces, etc. Vacuum processes are often also operated with supporting gases, which are required both actively as reactive gas or inactively as inert gas in the process. For this purpose, the gases are supplied to the process chamber 1 via valve 7, which is controlled by valve actuator 6 and can be used to control the gas supply and the pressure in process chamber 1. A measurement signal x is generated by the membrane pressure measurement cell 2, which is processed in the signal processing unit 4 and the control unit 5 to form a control signal s for the valve actuator 6. For precise process control, it is necessary that the membrane pressure measurement cell 2 measures as precisely as possible on the one hand, but also as quickly as possible on the other hand, in order to be able to react to pressure changes in the process chamber 1 as quickly and precisely as possible.

[0048] FIG. 1b shows a further simplified block diagram of a measurement cell assembly according to FIG. 1a, but now in a so-called “down-stream pressure control” instead of an “up-stream pressure control”. In the down-stream pressure control, the pressure gauge controls the conductance upstream of the vacuum pump via a controllable gas inlet valve 8. In contrast to the up-stream pressure control shown in FIG. 1a, in the down-stream pressure control shown in FIG. 1b, the vacuum pump 3 is connected to valve 7. Furthermore, the process chamber 1 is closed off via the controllable gas inlet valve 8. The gases required in the process chamber 1 are admitted to the process chamber 1 as required via the gas inlet valve 8.

[0049] It is also conceivable—in a simplified embodiment of the present invention—that the output signal y of the signal processing unit 4 is not used to control the pressure in a process chamber. It is then not a closed system, but an open system. In this case, a pressure in a pressure cell of any type—similar to the process chamber 1 according to FIG. 1—is measured with a pressure measurement cell 2. The measurement signal x measured with the pressure measurement cell 2 is also processed in a signal processing unit 4 to obtain a stable, noise-free output signal y which nevertheless reacts quickly to changes.

[0050] The invention now relates—again with regard to the embodiment variant according to FIG. 1—to the processing of the measurement signal x in the context of the conditions existing in a vacuum process and is primarily intended for optimum signal processing of the measurement signal x, as it can occur as a pressure signal in such vacuum processes. In this connection, the signal processing in the signal processing unit 4 can basically be carried out in an analog or digital manner, wherein the special precautions when signal processing is carried out in an analog manner or in a digital manner will not be discussed further in the following, since such precautions (analog/digital conversion, filtering to avoid aliasing, selection of the sampling frequency, etc.) are sufficiently known to the person skilled in the art.

[0051] The output signal y of the signal processing unit 4 is further processed in the control unit 5, for example with a so-called P, PI, PID or state controller. The controller implemented in the control unit 5 is responsible in particular for the optimum tracking of the control signal s for the valve actuator 6 or for the valve 7.

[0052] In principle, the explanations regarding the processes in the signal processing unit 4 and its block diagrams are valid both for the embodiment variant in a closed system and for the embodiment variant in an open system.

[0053] FIG. 2 shows in a schematic and simplified manner a block diagram of the processing steps according to the invention, which are processed in the signal processing unit 4. To implement the individual processing steps of the algorithm according to the invention, which is still to be explained, a signal processor is used, for example, which is programmed accordingly. It is understood that other tasks can also be performed by the signal processor, provided that the processor capacity is sufficient for this purpose. In particular, it is conceivable that the controller of the control unit 5 is also implemented in the same signal processor.

[0054] As can be seen from FIG. 2, the measurement signal x is fed to a first filter unit 10, which generates the output signal y. The first filter unit 10 with the measurement signal x and the output signal y form the actual signal path of the signal processing unit 4 (FIG. 1). The other components to be explained, such as a second filter unit 20 and an addition unit 11, are provided for defining the characteristics of the first filter unit 10.

[0055] The first filter unit 10 has a filter characteristic defined in a discrete-time system according to the following equation, for example:

y.sub.k=α.Math.x.sub.k+(1−α).Math.y.sub.k-1

[0056] Here, y.sub.k is the time-discrete output signal, x.sub.k is the time-discrete measurement signal, k is a time-dependent index, and α.sub.1 is a variable whose value decisively determines the time constant of the first filter unit 10 and is also referred to as the damping factor α.sub.1. The aim of the present invention is the optimum setting of the value for the damping factor α.sub.1, namely in such a way that a noise signal in the measurement signal x.sub.k is suppressed or even eliminated as far as possible, but at the same time a changing pressure in the process chamber 1 (FIG. 1) is quickly detected in order to be able to react to it correspondingly quickly.

[0057] The mentioned equation with the damping factor α.sub.1 has a low-pass characteristic as the filter characteristic for suppressing the noise signal component, wherein the time constant τ for a first-order filter at a sampling interval T can be determined as follows:

[00001]$\tau =T.Math.\frac{1-{\alpha }_1}{{\alpha }_1}$

[0058] The choice of values for the damping factor α.sub.1 is crucial for the present invention. If the measurement signal x.sub.k contains only a noise signal at a stable pressure value, the value for α.sub.1 is rather small, for example 0.0001. Thus, the noise signal present in the measurement signal x.sub.k is suppressed to a maximum and the filtered output signal y.sub.k is excellently suited for use in the downstream controller of the control unit 5 (FIG. 1), because a stable output signal leads to a lower activity of the valve actuator 6 or the valve 7 and thus to a reduced load on these components, which considerably reduces their probability of failure compared to known systems.

[0059] On the other hand, a change in the measurement signal x.sub.k due to an actual pressure change in the process chamber 1 (FIG. 1) is to be detected without delay, which necessitates a different value for the damping factor α.sub.1, namely, for example, a value for α.sub.1 greater than 0.3.

[0060] According to the invention, the value for the damping factor α.sub.1 is adjusted as a function of the difference between the output signal and the measurement signal. Starting from a discrete-time system in which the first filter unit 10 has a first-order low-pass filter according to the formula below,

y.sub.k=α.sub.1.Math.x.sub.k+(1−α.sub.1).Math.y.sub.k-1

the damping factor α.sub.1 is determined via an input difference x_diff or, in the discrete-time system, via x_diff.sub.k by determining—as can be seen from the analog system shown in FIG. 2—the difference between y.sub.k-1 and x.sub.k with the addition unit 11 as follows:

x_diff.sub.k=y.sub.k-1−x.sub.k

[0061] The input difference x_diff.sub.k is fed to the second filter unit 20, in which the first damping factor α.sub.1 is determined via a second damping factor α.sub.2. The second filter unit 20 again has, for example, first-order low-pass characteristics. Higher orders of low-pass filter characteristic are also conceivable. For first-order low-pass filter characteristics, in the case of a discrete-time system, the equation

α.sub.1k=α.sub.2.Math.x_diff.sub.k+(1−α.sub.2).Math.α.sub.1k-1

is applicable, wherein a second damping factor α.sub.2 is predefined. For example, the second damping factor α.sub.2 is in the range 0 to 1, more specifically in the range 0.05 to 0.25, even more specifically equal to 0.1.

[0062] Reference is made to the general fact that the damping factor α of a filter, in particular a first-order filter, can be expressed directly by the cutoff frequency f.sub.c and vice versa for those skilled in the technical field of filter design, whether in analog or discrete-time space. For a sampling interval T, the following formula is obtained for a first-order filter:

[00002]$\alpha =\frac{2.Math.\pi .Math.T.Math.f_c}{1+2.Math.\pi .Math.T.Math.f_c}$

or vice versa:

[00003]$f_c=\frac{\alpha }{2.Math.\pi .Math.T.Math.\left(1-\alpha \right)}$

[0063] This applies to both the first filter unit 10 and the second filter unit 20.

[0064] FIG. 3 shows the block diagram of a well-known exponential moving average filter of the first-order low-pass filter type as a discrete-time transfer function. The formula already mentioned applies:

y.sub.k=α.Math.x.sub.k+(1−α).Math.y.sub.k-1

wherein k is the index for time (and corresponding to k−1 a time delayed by a sampling interval T) and α is the damping factor.

[0065] Following the above formula, the block diagram shown in FIG. 3 is obtained with first and second adders 12, 13, a delay unit 15, and the damping unit 14, in which an output signal of the first adder 12 is multiplied by the damping factor α. An output signal of the damping unit 14 is applied to the second adder 13 in which it is summed with the delayed output signal y.sub.k-1 to produce the output signal y.sub.k. Finally, the output signal of the first adder 12 is formed by adding the input signal x.sub.k and the inverted delayed output signal y.sub.k-1.

[0066] The block diagram of the exponential moving average filter shown in FIG. 3 applies in principle to both the first filter unit 10 and the second filter unit 20.

[0067] FIG. 4 shows a further embodiment variant of the present invention, again using a block diagram. The measurement signal x.sub.k is again applied to a first filter unit 10 to generate the output signal y.sub.k. The first filter unit 10 again has first-order low-pass characteristics, although a higher-order filter can also be used.

[0068] As in the first embodiment variant of the present invention, which has been described with reference to FIG. 2, an addition unit 11 is provided in which the input difference x_diff.sub.k is generated by subtracting the input signal x.sub.k from the output signal y.sub.k-1. The input difference x_diff.sub.k is in turn applied to the second filter unit 20. The second filter unit 20 again has first-order low-pass characteristics, although a higher-order filter can also be used here.

[0069] The further embodiment variant of the invention shown in FIG. 4 is now, on the one hand, that the output difference y_diff.sub.k determined with the second filter unit 20 is applied to an absolute value unit 30 in which the absolute value of y_diff.sub.k, i.e. |y_diff.sub.k|, is determined. The absolute value |y_diff.sub.k| is applied to a functional unit 31 in which a function in the general form of a polynomial is applied to the absolute value |y_diff.sub.k|. A simplified function is, for example, the function implemented in the function unit 30 below:

z.sub.k=|y_diff.sub.k|.sup.exp

wherein an exponent exp is, for example, in the range 0 to 10, typically in the range 0.5 to 5, or even more typically equal to 2.5. It is self-evident that

z.sub.k=|y_diff.sub.k|

if exp=1, i.e. the embodiment variant according to FIG. 4 changes into the embodiment variant according to FIG. 2, in which the functional unit 31 can be regarded as not present.

[0070] The output value z.sub.k and a filter effect factor FW are fed to the multiplication unit 32, in which multiplication is performed to determine a product p.sub.k, which is fed to a limiting unit 33 for limiting to a value in the range 0 to 1. Thus, the first damping factor α.sub.1 of the first filter unit 10 is determined.

[0071] The function performed in the limiting unit 33 can be formally described as follows:

[00004]${\alpha }_1=\{\begin{array}{c}1,\mathrm{if}{z}_k.Math.\mathrm{FW}\ge 1\\ z_k.Math.\mathrm{FW},\mathrm{if}0<z_k.Math.\mathrm{FW}\le 1\\ 0,\mathrm{if}{z}_k.Math.\mathrm{FW}\le 0\end{array}$

[0072] It has been shown that the filter effect factor FW can be freely selected in the range from 0 to 10.sup.40.

[0073] Finally, the second damping factor α.sub.2 is chosen in the range of 0 to 1, typically in the range of 0.05 to 0.25, even more typically equal to 0.1.

[0074] Thus, the embodiment variant according to FIG. 4 comprises three predeterminable parameters: the filter effect factor FW, the second damping factor α.sub.2 and the exponent exp.

[0075] Of the three predeterminable parameters, the filter effect factor FW and the exponent exp are of particular importance. These two parameters have a decisive influence on the filter behavior: While the sensitivity of the filter can be adjusted via the exponent exp, the filter effect—as the name already expresses—can be adjusted via the filter effect factor FW, wherein the filter effect factor FW influences the noise component in the signal.

LIST OF REFERENCE SIGNS

[0076] 1 Process chamber [0077] 2 Membrane pressure cell [0078] 3 Vacuum pump [0079] 4 Signal processing unit [0080] 5 Control unit [0081] 6 Valve actuator [0082] 7 Valve [0083] 8 Gas inlet valve [0084] 10 First filter unit [0085] 11 Addition unit [0086] 12 First adder [0087] 13 Second adder [0088] 14 Damping unit [0089] 15 Delay unit [0090] 20 Second filter unit [0091] 30 Absolute value unit [0092] 31 Functional unit [0093] 32 Multiplication unit [0094] 33 Limiting unit [0095] x Measurement signal [0096] y Output signal [0097] s Control signal [0098] x_diff Input difference [0099] y_diff Output difference [0100] α Damping factor [0101] α.sub.1, α.sub.2 First and second damping factor [0102] FW Filter effect factor [0103] Exp Exponent [0104] z.sub.k Potentized output signal