DE102005029114B4 - Method for pressure measurement for combination pressure sensors in the vacuum range with at least two sensors and method for the automatic determination of an adjustment value for the "offset adjustment" in combination pressure sensors - Google Patents

Abstract

method for pressure measurement for combination pressure sensors in the vacuum area with at least two sensors, one each time-discrete, normalized, optionally linearized and / or adjusted for ambient temperature influences Sensor signal supply, with each adjacent in the pressure range Sensors have a common overlap area and to stabilize the output signal, a value adjustment carried out is characterized in that the range for the value adjustment depends on the course two or more sensor signals is determined.

Description

The The present invention relates to measuring pressure in a huge Range, in particular of absolute pressure in the range of atmospheric Pressure into the ultra-high vacuum range. There is no measuring principle is known that covers the entire area alone, are usually different Principles strung together.

to Saving costs and reducing space was Developed sensor arrangements, which combine several measuring principles.

For this Combination pressure sensors are methods of value matching in the Area of overlap from two different sensors necessary to even in this transition area to achieve a clear "output signal".

In The present invention describes a method in which by observing the signals of the participating sensors also outside of the overlap area Automatically determined a favorable "Wertangleichungsbereich" and thus the error in the output signal is reduced. Another method according to the invention, the same observation method The goal is to automatically readjust one of the pressure sensors in a state where its "sensor signal" is not included in the "output signal".

definitions

The following terms are used in the present specification:
A "sensor" is a sensor, consisting of sensor element and first signal conditioning, preferably in digital form, wherein microprocessor, A / D converter, time measurement with pulse width modulation or similar known means for digital signal processing can be used, optionally with supply unit.

Be Signal is called "sensor signal" and consists from normalized values, which are preferably already linearized and adjusted by ambient temperature influences are.

One such normalized "sensor signal" (f, g) has one Minimum resolution (Δf, Δg), which at a digital measured value output causes a maximum of one digital jump.

As the "output signal" A of the combination pressure sensor the signal is called the "sensor signals" of two or more Combined sensors may contain, preferably in digital Form, where optionally a further processing of the processed "sensor signal" into an analog signal he follows. The "output signal" A has a minimum resolution ΔA> Δf, Δg, otherwise no continuous value matching in the context of the dissolution is possible.

A graphical representation of the influencing variables is given in the 1 . 2 and 3 shown:

1 shows the true pressure for a first "sensor signal" ( 105 ) and the measured pressure ( 106 ) in logarithmic representation, as well as the ideal line ( 109 ).

The nominal characteristic of a real sensor is with ( 1015 ) designated.

Other real specimens of this type of sensor could be found in the area of tolerance lines ( 1011 ) and ( 1012 ) lie.

Likewise, a drift of the real sensor ( 101 ) to shifts within the tolerance lines ( 1011 ) and ( 1012 ) to lead.

The area between ( 1011 ) and ( 1012 ) is usually called a tolerance band. This tolerance band results in a "lower limit pressure" ( 1013 ) of the sensor ( 101 ).

2 shows the true pressure for a second "sensor signal" ( 205 ), the measured pressure ( 206 ), as well as the ideal line ( 209 ).

The nominal characteristic of a real sensor is with ( 2015 ) designated.

Other real specimens of this type of sensor could be found in the area of tolerance lines ( 2021 ) and ( 2022 ) lie.

In In the present application, these deviations are described as "sensor typical Tolerances ".

A drift of the real sensor ( 202 ) can cause shifts within the tolerance lines ( 2021 ) and ( 2022 ) to lead.

In In the present application, these deviations are referred to as "individual Tolerances ".

The area between ( 2021 ) and ( 2022 ) is again called the tolerance band. This tolerance band results in an "upper limit pressure" ( 2023 ) of this sensor.

3 shows an overlay 1 and 2 where the real sensors ( 301 ) and ( 302 ) not on the ideal line ( 309 ) are, as it often fig the practice corresponds.

( 3013 ) is the "lower limit pressure" of sensor ( 301 ) 3023 ) is the "upper limit pressure" of sensor ( 302 ). Between ( 3011 ) and ( 3012 ) as well as between ( 3021 ) and ( 3022 ) is the respective tolerance band.

As an "overlap area" ( 303 ), the range between "lower limit pressure" ( 3013 ) of the upper sensor ( 301 ) and "upper limit pressure" ( 3023 ) of the lower sensor ( 302 ) designated.

As "value adjustment area" ( 304 ), an area within the "overlap area" ( 303 ), in which the value adjustment takes place and which is defined dynamically according to the invention.

As a "lower observation area" ( 307 ), the area below the "lower limit pressure" ( 3013 ) of the upper sensor ( 301 ), as the "upper observation area" ( 308 ), the range above the "upper limit pressure" ( 3023 ) of the lower sensor ( 302 ) designated.

• Zustand 00: Es findet gerade keine Wertangleichung statt und das „Ausgabe-Signal" A entspricht dem „Sensor-Signal" g ((302), 3).
• Zustand 11: Es findet gerade keine Wertangleichung statt und das „Ausgabe-Signal" A entspricht dem „Sensor-Signal" f ((301), 3).
• Zustand 01: Die Wertangleichung findet ausgehend von „Sensor-Signal" g ((302), 3) in Richtung „Sensor-Signal" f ((301), 3) statt.
• Zustand 10: Die Wertangleichung findet ausgehend von „Sensor-Signal" f ((301), 3) in Richtung „Sensor-Signal" g ((302), 3) statt.
• ΔWA: „Wertangleichungsbereich".
• ε_p: „ Parallelitätstoleranz" (z.B. 2 Digits). Gibt an wie weit die Signale von der Parallelität abweichen dürfen um noch als parallel zu gelten.
• ε_s: „Stabilitätsabstand" (z.B. 10 Digits). Notwendig um den Gradienten der Signale sicher bestimmen zu können, damit zufällige Schwankungen (z.B. Digitsprung) nicht als Druckänderung interpretiert werden.
• N_min: minimale Anzahl von Schritten, die nötig sind um eine stetige Wertangleichung im Rahmen der Auflösung zu garantieren.

In the 4a . 4b . 4c the flowchart of the method for value matching is shown. Herein mean:

• State 00: There is currently no value matching and the "output signal" A corresponds to the "sensor signal" g (( 302 ) 3 ).
• State 11: There is currently no value adjustment and the "output signal" A corresponds to the "sensor signal" f (( 301 ) 3 ).
• State 01: The value matching takes place starting from "sensor signal" g (( 302 ) 3 ) in the direction of "sensor signal" f (( 301 ) 3 ) instead of.
• State 10: The value matching takes place starting from "sensor signal" f (( 301 ) 3 ) in the direction of "sensor signal" g (( 302 ) 3 ) instead of.
• ΔWA: "value adjustment range".
• ε _p : "Parallelism tolerance" (eg 2 digits) Indicates how far the signals may deviate from the parallelism in order to be considered parallel.
• ε _s : "Stability distance" (eg 10 digits) Necessary to be able to determine the gradient of the signals reliably, so that random fluctuations (eg digital jump) are not interpreted as a pressure change.
• N _min : minimum number of steps necessary to guarantee a constant value adjustment within the resolution.

State of the art

Already Developed in 1990, offered, sold and since then with the same In 1990 co-developed software are in operation with the systems Combination sensor VSKP 45M and the microprocessor measuring and control unit VD75 Company Thyracont Elektronic GmbH, Passau.

A public This system was presented at the 3rd Symposium Microsystem Technology on 17./18. February 1993 in Regensburg and at a lecture the industrial fair Achema in Frankfurt, held on June 7, 1994 as well as in the journal LABO 9/93.

In This combination pressure sensor is a piezoresistive absolute pressure sensor combined with a miniaturized Pirani sensor.

To overcome the discontinuity Both signals in a downstream electronics against each other weighted.

The Weighting is done in a linear manner, with a fixed End of the "overlap area" the "output signal" to 100% of the piezoresistive Sensor corresponds while at another specified lower point of the "overlap area", the "output signal" is 100% of that of the Pirani sensor corresponds.

So becomes a flowing transition created in the specified section of the "overlap area" and it creates a steady "output signal".

Reference is made to US Pat. No. 5,583,297. There above sensor VSKP 45M is called as state of the art. Its operation is obviously misinterpreted in this document, since that in the US 5 583 297 described method corresponds to the weighting described above in the evaluation of the sensor signals of the combination sensor VSKP 45M.

This known method has the disadvantage that in a predetermined Section of the "overlap area" the value matching is performed, what cause it can that the steady "output signal" a greater deviation from to have true value, as necessary.

If for example, the first sensor also in the "overlap area" high accuracy, So good agreement between measured and true value, the second sensor However, in the specified section or in a subset of the specified The section on value matching is flawed to a large extent, that will output continuous "output signal" by the percentage Inclusion of the faulty signal from the second sensor unnecessarily distorted.

An apparatus and method for pressure measurement that has two or more pressure senses is combined in the writing of the applicant DE 198 60 500 A1 described. Preferably, certain properties of a first sensor are used therein to correct deviations of a second sensor. A method for value matching is not apparent from this document.

In claim 17 of the US patent US 2002129657 describes a measuring method for sensor combinations with capacitive and Pirani sensor.

With the "sensor signal" of the second sensor we the offset in the "sensor signal" of the first sensor compensated. This is done by setting thresholds ("threshold values "), at this offset compensation takes place.

The German patent DE 44 45 534 C1 describes a method for obtaining an adjustment value for piezoelectric pressure sensors. It can not be used for pressure adjustment in the fine or high vacuum range.

Object of the invention

task The invention is a method for reducing measurement errors to describe with combination pressure sensors starting from the individual properties of the respective sensors the error automatically minimized in the "output signal".

• 1. eine Wertangleichung vorgenommen werden, die sich nur über einen tatsächlich notwendigen „Wertangleichungsbereich" (304) erstreckt und
• 2. eine Beobachtung des jeweils nicht in die Anzeige des Ausgabe-Signals einfließenden Sensors außerhalb des „Überlappungsbereiches" (303) erfolgen, um vorteilhaft bei Bedarf Justierwerte gewinnen zu können.

It should

• 1. an approximation of values can be made which only covers an actually necessary 304 ) and
• 2. an observation of the sensor not flowing into the display of the output signal outside of the "overlap area" (FIG. 303 ) in order to be able to advantageously obtain adjustment values as required.

The The first method should be able to automatically set the "value adjustment range" within the "overlap area" as small as possible during operation. This is to ensure be that the deviation of the output pressure value to the true Pressure is minimized, evaluated on all possible cases.

The Method should be applicable in any direction of pressure change.

The second method should be able to optionally using the method DE 198 60 500 A1 (see claims 14ff) to perform an automatic adjustment or readjustment.

This method should be suitable, for example, the zero pressure of a Pirani sensor or a capacitive sensor or a piezoresistive sensor or other sensors for the coarse and fine vacuum range by observing the relevant sensors on the "overlap" ( 303 ) in the "lower observation area" ( 307 ).

Furthermore, the method should be suitable, for example the system pressure of a capacitive sensor or the upper measuring range limit of other sensors for vacuum by observing the relevant sensors via the "overlap area" (US Pat. 303 ) in the "upper observation area" ( 308 ).

The Observing the sensors via The "overlap area" should also be to a statement about the current functionality the sensors can be used.

Inventive solutions

The solutions according to the invention are based on a pairwise overlapping of several pressure ranges (see FIG. 3 ).

The signal f of a sensor ( 301 ) corresponds to the upper pressure range, the signal g of a second sensor ( 302 ) the lower pressure range.

For the sensor ( 301 ) there is a "lower limit pressure" ( 3013 ) f _Min , which defines the "overlap area" ( 303 ) closes down.

For the sensor ( 302 ) there is an "upper limit pressure" ( 3023 ) g _Max , which defines the "overlap area" ( 303 ) closes upwards.

Both Sensors provide a signal that follows the true pressure, but within the tolerance lines may differ from the true value.

solution for task 1 (value adjustment):

The inventive solution of Task 1 is based on an automatic determination of the value adjustment range.

Therefore should first be discussed briefly be when it is cheap is to start with the alignment of values and why these in one preferably small area should be done.

Within the "overlap area" ( 303 ) Consider the function Δ (p) = log b (f (p)) - log b (G (p)), as a function of the true pressure p.

Are the characteristic curves of the sensors in the double logarithmic diagram ( 3 ) in parallel, Δ (p) = C is constant and it holds that c = b ^C : f (p) = c g (p), that is, the signals f and g differ by the constant factor c, the z. B. may be the gas type correction factor of a Pirani sensor.

run the characteristic curves of the sensors in the double logarithmic diagram not parallel, Δ (p) grows locally either monotonously or Δ (p) falls monotonously, because the characteristics are monotone functions. But that means that the signals f and g either run towards each other or themselves remove each other.

The quotient of the signals f and g changes with the exponential function b ^{Δ (p)} : f (p) / g (p) = b Δ (p) ,

If, for example, in such a region in which Δ (p) grows monotonically or monotonically, the value equalization takes place, for example, with a linear percentage weighting of the "sensor signals", a deviation of the "output signal" from the true value occurs, which is unnecessary high is:
Due to the drifting apart of the "sensor signals" f and g in accordance with the exponential function b ^{Δ (p)} , at least one of the "sensor signals" must deviate from the true value to an increasing or decreasing extent.

This Error then enters the linearization of the value and leads to unnecessarily large deviations.

finds However, the value adjustment takes place in a range in which Δ (p) in the frame the measurement accuracy is constant or approximately constant, then soft the "sensor signals" only by a constant Factor off each other. This falsifies with a constant value approximation no "sensor signal" the other one more as necessary.

Farther the alignment should take place in the smallest possible area. Since the characteristic curve is not known a priori, there is the Possibility, that the "sensor signals" start apart to drift and the value adjustment would then in the unfavorable Area will be continued.

Below is the procedure for matching the values using the flow charts 4a -c are described.

The diagram in 4a shows the general flow chart for the value adjustment:
First, an error and plausibility check ( 401 ) checks whether a value adjustment makes sense (eg a sensor could be defective). If there is no error, the procedure ( 402 ) to be continued.

The process is initially divided into two large blocks. Namely the subroutine "find value adjustment area" ( 404 ), if state = 00 or state = 11, and the subroutine "value adjustment is running" ( 403 ) when the other two states 01 and 10 are present.

Based on the flowchart in 4b Now the subroutine "find value matching range" is described:
It serves to find a criterion for the beginning of the value adjustment and to determine the size of the "value adjustment range".

To Condition is either state = 00 or state = 11, i.e. it finds by definition just no value adjustment takes place.

First, it is now checked in which area the "sensor signals" are located ( 412 ) 413 ).

Is the signal f of the sensor ( 301 . 3 ) above the "overlap area" ( 303 ), f is displayed as "output signal" A and state = 11 ( 411 ), but if it is below, the signal g of the sensor ( 302 ) and state = 00 ( 415 ) set. The current signals are stored (f ₀ = f, g ₀ = g) ( 421 ) 422 ).

As can be seen in the flowchart in 4b detects, the process now disintegrates depending on the state ( 414 ) in two branches ( 416 ) 431 ) and there is at least one "sensor signal" f or g within the "overlap area".

• Zustand = 11, fallender Druck → Wertangleichung von f nach g.
• Zustand = 00, steigender Druck – Wertangleichung von g nach f.

The state determines in which direction an eventually occurring value adjustment should take place:

• State = 11, falling pressure → value adjustment from f to g.
• Condition = 00, increasing pressure - value adjustment from g to f.

For further description of the method oBdA should now be assumed that the "sensor signal" f ( 301 ) within the "overlap area" ( 303 ) and state = 11 exists (branch of ( 414 ) to ( 416 ), left part of the flowchart in 4b ).

It is now checked whether the "sensor signal" g within the "overlap area" ( 303 ) lies ( 416 ). Is that the case (( 416 ) "No"), it must be checked, because of the stability of the method, whether the "sensor signal" f is at least equal to "Stability distance" ε _{s is} smaller than the previously stored "sensor signal" f ₀ ( 417 ).

If this is true, then the following terms are calculated ( 418 ): Δ 0 = | log b (f 0 ) - log b (G 0 ) |, Δ 1 = | log b (flew b (G) |, N min = Δ 1 / (ΔA - max (Δf, Δg)

Δ ₀ and Δ ₁ describe the difference of the signals on a logarithmic scale, N _min is the minimum number of value matching steps necessary within the resolution for a continuous value equalization.

On the basis of the values, it is now necessary to check whether the conditions for the beginning of a value approximation are given ( 419 ). The following cases are to be distinguished: 1.) If f - f _min ≦ N _min · max (Δf, Δg), then the value equalization must be started immediately in order to achieve a steady value equalization within the resolution max (Δf, Δg) without falling below the lower limit pressure f _min (( 419 ) "Yes").

• 2.) Ist |Δ₁ – Δ₀| ≤ ε_p ((425) „nein"), dann unterscheiden sich die „Sensor-Signale" um weniger als die „Parallelitätstoleranz" ε_p, d.h. sie verlaufen im doppelt logarithmischen Diagramm annähernd parallel und die „Sensor-Signale" f und g unterscheiden sich um einen konstanten Faktor. Es wird daher mit der Wertangleichung begonnen (427). Das Intervall des „Wertangleichungsbereichs" (304) wird nun nach der Formel ΔWA = Nmin·max(Δf, Δg)dynamisch bestimmt (427). Mit diesem Intervall ist gewährleistet, dass die Wertangleichung auf einen Bereich beschränkt bleibt, der der geforderten Auflösung entspricht. Die obere Grenze des „Wertangleichungsbereichs" (304) ist f₀, die untere Grenze berechnet sich zu f₀ – ΔWA. Wie in Fall 1.) werden die „Sensor-Signale gespeichert (427) und der Zustand = 10 (436) gesetzt.
• 3.) Ist Δ₁ – Δ₀ > ε_p ((425) „nein"), dann muß sofort mit der Wertangleichung begonnen werden, denn die „Sensor-Signale" f und g beginnen auseinander zu laufen. Die Speicherung, das Setzen des Zustandes und des „Wertangleichungsbereichs" erfolgt genauso wie in Fall 2.).
• 4.) Ist Δ₀ – Δ₁ > ε_p, dann laufen die „Sensor-Signale" f und g noch aufeinander zu und es gibt nichts zu tun ((425) „ja").

ΔWA = f - f _{min is} set as the interval of the "value adjustment range", with f ₀ : = f as the upper limit and f _min = f ₀ - ΔWA as the lower limit ( 420 ). It becomes the state = 10 ( 436 ) (Value adjustment from f to g running) is set and the current "sensor signals" (f ₀ : = f, g ₀ : = g) are stored ( 420 ). If the above condition is not fulfilled, then go to case 2.) (( 419 ) "No"):

• 2.) Is | Δ ₁ - Δ ₀ | ≤ ε _p (( 425 ) "No"), then the "sensor signals" differ less than the "parallelism tolerance" ε _p , ie they run approximately parallel in the double logarithmic diagram and the "sensor signals" f and g differ by a constant factor , It is therefore begun with the value adjustment ( 427 ). The interval of the "value adjustment range" ( 304 ) will now follow the formula ΔWA = N min Max (Δf, Δg) determined dynamically ( 427 ). This interval ensures that the value adjustment is limited to an area that corresponds to the required resolution. The upper limit of the "value adjustment range" ( 304 ) f ₀ , the lower limit is calculated as f ₀ - ΔWA. As in case 1.) the "sensor signals are stored ( 427 ) and the state = 10 ( 436 ) set.
• 3.) Is Δ ₁ - Δ ₀ > ε _p (( 425 ) "No"), then the value adjustment must be started immediately, because the "sensor signals" f and g start to diverge. The storage, the setting of the state and the "value adjustment range" are the same as in case 2.).
• 4.) If Δ ₀ - Δ ₁ > ε _p , then the "sensor signals" f and g still converge and there is nothing to do (( 425 ) "Yes").

The cases 2.), 3.) and 4.) are in the flow chart in 4b to a case distinction (Δ ₀ - Δ ₁ > ε _p ?) ( 425 ), since cases 2.) and 3.) do not differ in their execution.

Analogous to the method just described, in the right part of the flowchart in FIG 4b if the "sensor signal" g ( 302 ) within the "overlap area" ( 303 ) lies ( 413 ) and state = 00 ( 414 ) is present.

• Zustand = 10, fallender Druck → Wertangleichung von f nach g.
• Zustand = 01, steigender Druck → Wertangleichung von g nach f.

The description of the find value matching subroutine is now complete and will now be described in the flowchart in FIG 4c the subprogram "Value Matching" will be explained:
By definition, the system is in one of the states 10 or 01 ( 441 ). Hereby means:

• State = 10, falling pressure → value adjustment from f to g.
• State = 01, increasing pressure → value adjustment from g to f.

First will now check if yourself the "sensor signal" f still within of the "value adjustment area".

Is the "sensor signal" f above the "value adjustment range"(f> f ₀ : state = 10, f> f ₀ + ΔWA: state = 01) (( 443 ) respectively. ( 442 ) "Yes"), state = 11 is set and the "sensor signal" f is supplied to the "output signal" A ( 445 ).

Is the "sensor signal" f below the "value adjustment range" (f <f ₀ : state = 01, f <f ₀ - ΔWA: state = 10) (( 446 ) respectively. ( 447 ) "Yes"), state = 00 is set and the "sensor signal" g is supplied to the "output signal" A ( 451 ).

In both cases, the current "sensor signals" (f ₀ : = f, g ₀ : = g) are stored ( 444 ) 448 ).

Is the "sensor signal" f within the "value adjustment range" ( 304 ) (( 446 ) respectively. ( 447 ) "No"), an equation of value is found according to the following formula ( 452 ) instead of: A (f, g) = a (f) · f + (1-a (f)) · g, where A corresponds to the "output signal" and a (f) is a continuous weighting function.

Depending on the state (value matching of f to g or g to f), the weighting function a (f) is now set:
For example, a (f) can be the linear function.
( 450 ) a (f) = 1 - (f ₀ - f) / ΔWA: state = 10 or
( 449 ) a (f) = (f - f ₀ ) / ΔWA: state = 01
be.

At the "upper limit" of the dynamically determined "value adjustment area" ( 304 ), the "output signal" corresponds to 100% of the "sensor signal" f of the first sensor ( 301 ), at the "lower limit" to 100% of the "sensor signal" g of the second sensor ( 302 ).

In order to Now the subroutine "Value Matching is running" is completely described.

solution for task 2 (obtaining calibration values):

Prerequisite for this method is a digital signal processing preferably by means of A / D converter (ADC = analog to digital converter or TDC = time to digital converter, optional as in DE 101 15 715 B4 ) and microprocessor.

The inventive method is suitable to run fully automatically.

It However, it can also be started specifically.

Example 1:

The sensor to be adjusted is ( 301 ).

Is the "sensor signal" g of the sensor ( 302 ) within the "lower observation area" ( 307 ), so the distance of this signal is continuously (or at short intervals) to the "lower limit pressure" f _min ( 3013 ) detected.

When the "sensor signal" g reaches a value that, for example, two decades below the "lower limit pressure" f _Min ( 3013 ), an observation timer is started.

If, within an observation time set for this timer, the "sensor signal" f of the sensor to be adjusted ( 301 ) is stable, that is, when the value of the "sensor signal" f of the sensor to be adjusted, for example, no more than +/- 2 A / D steps or time steps changes and if at the expiration time of the observation time, the deviation of the current "Sensor signal" f from the last "offset" adjustment value (f) exceeds a predetermined maximum allowable amount (preferably some A / D steps or time steps),
then the value of this currently applied "sensor signal" f is automatically stored as a new "Offset" -Justierwert for the sensor to be adjusted.

Example 2:

The sensor to be adjusted is ( 302 ).

Is the "sensor signal" f of the sensor ( 301 ) within the "upper observation area" ( 308 ), the distance of this signal to the "upper limit pressure" g _max (or at short intervals) is continuously ( 3023 ) detected.

If the "sensor signal" f reaches a value, for example, a decade above the "upper limit pressure" g _max ( 3023 ), an observation timer is started.

If within a given observation time the "sensor signal" g of the sensor to be adjusted ( 302 ) is stable, that is, if the value of the "sensor signal" g of the sensor to be adjusted, for example, no more than +/- 2 A / D steps or time steps changes and if at the expiration time of the observation time, the deviation of the current "Sensor signal" g from the last "offset" adjustment value (g) exceeds a predetermined maximum allowable amount (preferably some A / D steps or time steps),
then the value of this currently applied "sensor signal" g is automatically stored as a new "offset" -Justierwert for the sensor to be adjusted.

Example of self-diagnosis:

If the deviations of the current, current "sensor signals" from the last stored calibration values can deviate seriously Defective on the respective sensor element or on the relevant measuring chain getting closed. Optionally, the output of a noise signal respectively.

Achieved benefits:

in the In contrast to the method of measured value matching according to U.S. Pat. patent No. 5,583,297, which operates on a previously established "value adjustment range", will become the Starting point of the adjustment according to the invention dynamically as a function of calculated from the output signals of the sensors. This prevents that an alignment takes place already when the sensor signal a sensor is still very inaccurate and thus a falsification of the output signal.

If, due to the influence of different measuring media, aging or other effects flows changes the characteristic of one of the sensors, both the automatic value adjustment method according to the invention and the method according to the invention for adjustment will result in an improvement in the accuracy of the output signal.

At the Example of a Pirani sensor can be shown by the dynamically determining the area for value matching the approximation in Linear range takes place where the accuracy of the sensor on highest is.

By Method 2 ensures that the characteristics of the real Sensors always as close as possible on the ideal line.

This has a reduction of the error in the output signal even outside of the "Wertangleichungsbereiches" result.

The inventive method are able to engage themselves by incorporating the observation areas both regarding the sensor typical as well as the individual sensor characteristics automatically adjust and thus reduce errors in the output signal, as well as indications of disturbances to give.

Legend: 4b

L001:

Storage of the current values, f ₀ = f, g ₀ = g

L002:

f <f ₀ - ε _s , ε _s eg 10 digits

L003:

g> g ₀ + ε _s , ε _s eg 10 digits

L004:

Δ ₀ = | log _b (f ₀ ) - log _b (g ₀ ) |, Δ ₁ = | log _b (f) - log _b (g) |,

N _min = Δ ₁ / (ΔA - max (Δf, Δg))

L005:

Δ ₀ - Δ ₁ > ε _p , ε _p eg 2 digits

L006:

f -f _min <N _min -max (Δf, Δg)

L007:

g _max -g <N _min -max (Δf, Δg)

L008:

Storage of the values, f ₀ = f, g ₀ = g,

ΔWA = N _min × max (Δf, Δg)

L009:

Storage of the values, f ₀ = f, g ₀ = g,

ΔWA = f - f _min

L010:

Storage of the values, f ₀ = f, g ₀ = g,

ΔWA = g _max - g

Claims (14)

Method for pressure measurement for combination pressure sensors in the vacuum range with at least two sensors, each of which provides a time-discrete, normalized, optionally linearized and / or adjusted for ambient temperature influences sensor signal, wherein each adjacent sensors in the pressure region have a common overlap region and for the stabilization of the output signal, a value adjustment is carried out, characterized in that the range for the value equalization is determined depending on the course of two or more sensor signals. Method according to claim 1, characterized in that that the extension of the range for the value adjustment is so small as possible chosen to achieve a reduction of the measurement error in the output signal. Method according to claim 1 or 2, characterized that it is applicable in both directions of a pressure change. Method according to one of Claims 1 to 3, with the following steps, step 1 - First, an "error and plausibility check" ( 401 ): monitoring of both sensors for breakage and short circuit. Monitoring the deviation of both sensors to their last adjustment values and the deviation of the sensor signals to each other. Step 2 - If no value adjustment is taking place, then the "output signal" A corresponds to the sensor signal "g (state = 00) or the" sensor signal "f (state = 11) (( 402 ) 4a ). It is with step 4 "find value matching area" ( 404 ). Step 3 - A value adjustment takes place starting from "sensor signal" f in direction "sensor signal" g (state = 10) or from "sensor signal" g in direction "sensor signal" f (state = 01) (( 402 ) 4a ). It starts with step 5 "Value Matching" ( 403 ). Step 4 - In the step "Find Value Matching Area" ( 404 ) is checked whether the conditions for the beginning of the value adjustment are given and if so, the value adjustment range is determined depending on the course of one or more sensor signals. Step 5 - In the step "Value Matching" ( 403 ), the approximation of the values is within the specified value adjustment range ( 304 ) carried out. Method according to one of Claims 1 to 4, characterized in that - the interval ΔWA of the value matching range ( 304 ) by the following formulas ( 418 ) 433 ) is determined: Δ 0 = | log b (f 0 ) - log b (G 0 ) |, Δ 1 = | log b (flew b (G) |, N min = Δ 1 / (ΔA - max (Δf, Δg), If the minimum distance ( 419 ) 434 ) N _min · max (Δf, Δg) to the "lower limit pressure" ( 3013 ) or "upper limit pressure" ( 3023 ) below: ΔWA = f - f _min ( 420 ) or ΔWA = g _max - g ( 435 ) In all other cases: ΔWA = N _min × max (Δf, Δg) ( 427 ) 428 ) Method according to one of claims 1 to 5, characterized in that the implementation of the step "value adjustment is running" ( 403 ) using the following formulas and instructions: The "output signal" A follows the formula: ( 452 A (f, g) = a (f) * f + (1-a (f)) * g where f is the "sensor signal" ( 301 ) and g the "sensor signal" ( 302 ) and wherein the weighting function a (f) for the complete run of a value match is determined as follows: ( 450 ) a (f) = 1 - (f ₀ - f) / ΔWA, when the value is equalized from f to g and ( 449 ) a (f) = (f - f ₀ ) / ΔWA, when the value is equalized from g to f. - The value adjustment takes place as long as the "sensor signal" f is within the "value adjustment range". Method for the automatic determination of an adjustment value for the "offset adjustment" in combination pressure sensors with the following features: The upper pressure range and the lower pressure range have an "overlap range" ( 303 ), - Below the "overlap area" ( 303 ) closes the "lower observation area" ( 307 ), - Above the "overlap area" ( 303 ) closes the "upper observation area" ( 308 ), and the following steps: Receiving the "sensor signal" ( 301 ) of a first pressure sensor suitable for measuring pressure in an upper pressure range, - receiving the "sensor signal" ( 302 ) of a second pressure sensor suitable for pressure measurement in a lower pressure range, characterized in that a new adjustment value for the "offset adjustment" is found if a) the "sensor signal" ( 302 ) within the "lower observation area" ( 307 ), wherein it is continuously monitored and - when falling below a sensor signal value ( 302 ) at a defined distance below the "lower limit pressure" ( 3013 ), an observation timer (time "tb") is started, - the "sensor signal" ( 301 ) does not fluctuate by more than a predetermined amount during the observation time "tb", the observation timer (tb) is restarted if the fluctuation is above the predetermined amount - the deviation of the "sensor signal" ( 301 ) exceeds a defined maximum permissible value from the last adjustment value, b) the "sensor signal" ( 301 ) within the "upper observation area" ( 308 ), wherein it is continuously monitored and - when a sensor signal value ( 301 ) at a defined distance above the "upper limit pressure" ( 3023 ), an observation timer (time "tb") is started, - the "sensor signal" ( 302 ) does not fluctuate by more than a predetermined amount during the observation time "tb", the observation timer (tb) is restarted if the fluctuation is above the predetermined amount - the deviation of the "sensor signal" ( 302 ) exceeds a defined maximum allowable amount from the last adjustment value. Method according to one of the preceding claims, characterized in that an error signal is output when the "sensor signals" ( 301 ) and ( 302 ) are outside a predefined tolerance band. Method according to one of the preceding claims characterized characterized in that single, multiple or all sensor signals a combination sensor, currently not included in the output signal will also be over the overlap area with other sensors. Method according to one of the preceding claims characterized characterized in that the "sensor signal" of the first sensor from a piezoresistive sensor and the "sensor signal" of the second sensor from a Pirani sensor. Method according to one of the preceding claims characterized characterized in that the "sensor signal" of the first sensor from a capacitive membrane sensor and the "sensor signal" of the second sensor from a Pirani sensor comes. A method according to any preceding claim characterized in that the "sensor signal" of the first sensor from a Pirani sensor and the "sensor signal" of the second sensor from a sensor with an electron source, in particular a hot, cold, Feldüberhöhungskathode comes. Method according to one of the preceding claims, characterized in that the "sensor signal" of the first sensor of a capacitive membrane sensor and the "sensor signal" of second sensor from a sensor with an electron source, in particular a hot, cold, Feldüberhöhungskathode originated. Method according to one of the preceding claims characterized characterized in that the "sensor signal" of the first sensor from a capacitive membrane sensor and the "sensor signal" of the second sensor from a capacitive membrane sensor.