WO2011032763A2

WO2011032763A2 - Transducer stage for detecting a sensor signal

Info

Publication number: WO2011032763A2
Application number: PCT/EP2010/060869
Authority: WO
Inventors: Dayo Oshinubi; Alexander Buhmann
Original assignee: Robert Bosch Gmbh
Priority date: 2009-09-15
Filing date: 2010-07-27
Publication date: 2011-03-24
Also published as: DE102009029481A1; WO2011032763A3

Abstract

The invention relates to a transducer stage for detecting a sensor signal, in particular a signal of a rotational rate sensor, comprising an amplification stage with an electronic circuit and characterised in that various weighing functions can be produced by means of said electronic circuit, which can influence a filter characteristic of the transducer stage. Said claimed transducer stage has the advantage that, compared to traditional systems, the noise outlet intensity in the base band is reduced when sampling and sub-sampling said sensor signal.

Description

description

title

Transducer stage for detecting a sensor signal The invention relates to a converter stage and a method for detecting a

Sensor signal, in particular a signal of a rotation rate sensor.

PRIOR ART In micromechanical rotation rate sensors, time-continuous sensor signals are converted by an analog / digital conversion (A / D conversion) into discrete-time and discrete-value signals in order to carry out signal processing such as filtering, regulation and further algorithms within a digital circuit part. The digitization of the sensor signals plays a decisive role, since this operation decisively influences the noise behavior of the entire system.

A sensor system with a digital evaluation concept provides capacitance / voltage conversion (C / U conversion), in particular for capacitive sensors, in order to convert capacitance changes caused by external forces into proportional electrical voltages. In time discrete C / U conversion, sampling is conventionally done at very high frequencies. Common are scans of 8 to 40 times compared to a sensor resonance frequency. Due to these high sampling rates, an analog front-end device has a very high electrical power requirement. In many areas, this oversampling of the sensor signal makes sense, since in this way a reduction of the noise effects by aliasing can be achieved.

However, for example, in mobile applications, the current consumption of an overall system is very important because it depends on a degree of operational readiness of the application. Therefore, in such applications mostly a compromise between sampling rate and noise performance can be found. A reduction of the sampling rate for the purpose of a power reduction, however, would mean a significant increase in the noise power density due to noise folding within the baseband in the signal evaluation.

From DE 10 2007 048 825 A1, an inertial sensor, in particular a rotation rate sensor for detecting at least one parameter predetermined by a drive element, in particular rotation rate and / or rotational speed as a function of an input signal to be scanned, is known. It is provided that an undersampling of the input signal takes place at a sampling frequency which is below a double maximum frequency of the input signal. Disadvantageously, due to this undersampling, a contribution may be increased by noise convolution into the baseband.

It is therefore an object of the invention to provide a device for detecting sensor signals, which does not have the above-mentioned disadvantages. Disclosure of the invention

The circuit according to the invention is designed as a converter stage for detecting a sensor signal, in particular a signal of a rotation rate sensor, and comprises an amplifier stage with an electronic circuit. The converter stage is characterized in that different weight functions can be generated by means of the electronic circuit with which a filter characteristic of the converter stage can be influenced.

The converter stage according to the invention with the features mentioned offers the advantage that it can be used to achieve a reduction of the noise power density in the baseband during sampling and undersampling of the input signal compared to conventional systems. Furthermore, by means of the influenceable filter characteristic, a transmission behavior of the converter stage can advantageously be configured in such a way that, in the case of sampled, subsampled and bandpass sampled systems, a contribution to the baseband (aliasing of the

Broadband noise) is reduced. This reduction makes it possible to To optimize velvet performance in such sensor system realizations. Advantageously, this can for example minimize energy consumption of the overall system and, due to this minimization, additional system components, for example an I / Q demodulator, can be integrated into a single front-end device. Furthermore, as a result, an ASIC chip area can also advantageously be reduced, which is advantageous, in particular, in the case of multiaxial sensor applications with a plurality of sensors with an increased chip area requirement. Both sampling, filtering and demodulation are thus advantageously possible within a single frontend.

According to a preferred embodiment, the converter stage comprises a controllable current source and an electrical charge integrator, wherein the electrical charge integrator can be driven by an output of the amplifier stage. As a result of this interaction between the amplifier stage and the electrical charge integrator, a filter characteristic of the overall system can be controlled in a simple manner, so that an image of the sensor signal can be tapped off as an electrical voltage at the charge integrator.

According to a further preferred embodiment, the amplifier stage is designed as a transconductance amplifier, wherein the current source is designed as Biasstrom- source for the transconductance amplifier. Thereby, the bias current source, which is usually required for setting operating points for each transconductance amplifier, can usefully be used additionally for adjustment of the weight function, which advantageously requires no additional electronic circuitry for generating the weight function.

According to a further preferred embodiment, the current source can be driven by a rectangular control signal transmitted via a control line. As a result, a simple window or weight function for the input current signal of the amplifier stage can be realized by trivial switching on and off of the current source.

A further preferred embodiment of the invention provides that the control signal periodic switching cycles with N substantially equal intervals having. By specifying the individual switching cycles, a multiplicity of different filter characteristics can be realized in this way.

A further preferred embodiment provides that the current source can alternatively be driven by a control signal transmitted via a control line rectangular, triangular, or in accordance with a Hanningfunktion control signal, the control signal having periodic switching cycles with N substantially equal intervals. This corresponds to an application of pulsed weight functions with each switching cycle differently ausgestaltetbarer waveform. With this variant, the advantages of other variants of the invention can be combined and different filter characteristics can be achieved in even greater variety with even improved suppression of interference spectra.

According to an advantageous development of the invention, an integration time constant of the electrical charge integrator is variably adjustable. A parametrization of a filter characteristic of the converter stage according to the invention can be carried out in this way simply by varying the integration time constant.

Brief description of the drawings

The invention will be briefly described below with reference to figures. Showing:

Figure 1 is a general block diagram of a sensor system in which the

Invention can be used

Figure 2 is a simplified electrical diagram of a converter stage according to the invention

Figure 3 shows an application of different weight functions and resulting transmission behavior

Figure 4 rectangular time course of the bias current Figure 5 shows a transmission behavior of the converter stage according to the invention with a pulsed rectangular bias current

FIG. 6 Timing curves of the bias current with triangular and Hanningfenste- tion

FIG. 7 shows a frequency spectrum with pulsed triangular window operation of the inventive converter stage embodiment (s) of the invention

Figure 1 shows in a general block diagram a sensor system 10 with a field of application of the present invention. The sensor system 10 has a sensor 20 which is galvanically connected to a C / U converter 30 and whose output signal in the form of a capacitance is fed to the C / U converter 30. An output signal of the C / U converter 30 is supplied to a converter stage 40. In FIG. 1, an area of application of the present invention is emphasized by a dot-dashed border. Within this range, the conversion of the capacitance signals into proportional electrical voltages and a sampling of these sensor detection signals take place. The further structure and the further function of the sensor system 10 with the remaining components sample / hold element 50, quantizer 60, digital circuit part 70 and digital / analog converter 80 will not be further explained below, since these components are not essential to the present invention and in general represent known elements of a sensor system.

FIG. 2 shows, in greater detail, a simplified electrical circuit diagram of a converter stage 40 according to the invention. The converter stage 40 assumes the function of a digital / analog filter and comprises an electrical amplifier stage 41, which is galvanically connected to a bias current source. The amplifier stage 41 is preferably designed as a transconductance amplifier and is driven by the bias current source 42, whereby the amplifier stage 41 is controllable in its transmission behavior. The transconductance amplifier is a specific operational amplifier which converts a differential voltage Vin at its inputs into a proportional output current Iout. An output of the amplifier stage 41 controls with this output current lout an electrical charge integrator galvanically connected to this output. The charge integrator comprises a first switch 43, a second switch 46 and a capacitor 44, to which an output voltage Vout of the converter stage 40 can be tapped off as an image of a sensor signal of the sensor 20. An operating behavior of the bias current source 42 can be controlled by means of a control line 45, via which different control signals for the bias current source 42 are transmitted. As a result, a current flow of the bias current source 42 and, as a result, a transmission behavior of the converter stage 40 can be influenced. In other words, by means of the controllable bias current source 42, output currents lout of the amplifier stage 41, which correspond to input currents lin of the charge integrator, are weighted with different weight functions. By switching the two switches 43, 46, the current I _{out is} switched on the one hand and furthermore the capacitor 44 is charged with the switched current I out in accordance with a switching pattern of the switches 43, 46. In this case, an integration behavior of the electrical charge integrator can be changed by varying the switching times. As a result of the integration, an output voltage V _out is available on the capacitor 44 as an image of the sensor signal of the sensor 20. The electrical charge integrator is resettable by means of the second switch 46, so that its operating characteristics can be initialized after each reset.

The controllable bias current source 42 thus makes it possible in the above-described manner to weight the input currents lin for the charge integrator with different weight functions. For the output voltage Vout of the converter stage 40, this results in the following mathematical relationship:

, -ίί

Vsmt = j ta iffi- W t

In it mean: t time

V _out output voltage of the converter stage 40

to, t1 integration period

W (t) weight function iout output current of the amplifier stage 41 (input current of the charge integrator)

The output voltage V _out thus represents a time integral of the weighted output current lout of the amplifier stage 41 or input current of the charge integrator. By simply changing the switching times tO, t1 of the switches 43, 46 and thus the integration time of the charge integrator, the transmission behavior and thus the output signal can be the converter stage 40 vary slightly.

FIG. 3 qualitatively shows a transmission behavior of the converter stage 40 when different weight functions (window functions) W are used. In the left-hand diagram of FIG. 3, three examples of different control signals, with which the bias current source 42 is actuated via the control line 45, are shown in a time range. This corresponds to three different weight functions rectangle, triangle and Hanningfunktion, which can be generated by each specific control of the bias current source 42 with the different control signals. Instead of the illustrated signal forms of the control signal, other signal forms can also be used. The right-hand diagram of FIG. 3 qualitatively shows, in a different frequency range, how differently the three aforementioned control signals affect a transmission behavior of the converter stage 40. It can be clearly seen that with weight functions W of higher order than the rectangular function (such as, for example, triangular function or Hanning function), greater attenuation towards higher frequencies can be achieved. The greatest damping can be achieved with the function function W in terms of weight. This is equivalent to an improved suppression of aliasing noise spectra within the baseband.

A further embodiment of the invention provides for a subdivision of the sampling interval into N substantially equal time intervals, so that the following mathematical relationship applies: By segmenting into arbitrary time intervals into which the control signal for the bias current source 42 can be divided, a multiplicity of different filter characteristics can be generated for the converter stage 40. Figure 4 shows a time principle of such a segmentation of the control signal, which can be realized by simply switching on and off of the bias current source 42 via the control line 45. With such a drive, different switching patterns for the bias current source 42 can be generated. In the example of FIG. 4, a switching pattern {1, 0, -1} is shown over the three switching intervals I, II and III. The symbol "1" corresponds to a positive bias current I _B i _as . The symbol "-1" corresponds to a negative _bias current I _bias . The timing diagram of FIG. 4 shows a time characteristic of the bias current IBias with a total sampling time TA, which is essentially subdivided into N equal switching intervals Ti denoted by I, II and III. In the switching interval I, the bias current I _B i _{as has} a constant positive value. In a subsequent switching interval II, the bias current I _B i _as is substantially equal to zero and in a subsequent switching interval III the bias current I _B i _{as has} a constant negative value. All of these switching patterns are transmitted to the bias current source 42 via control line 45. The switching cycles can be repeated as desired, so that such a transmission behavior of the converter stage 40 can be achieved, which is characterized by a suppression of higher and lower frequency frequency components, which is synonymous with a favorable noise shaping for the sensor system 10th

FIG. 5 qualitatively shows four different frequency spectra of the transducer stages 40 which result from four different switching patterns. The transmission characteristics shown in FIG. 5 result from a pulsed rectangular control signal for the bias current source 42 with the following four different switching patterns shown from top to bottom: {-1, 0, 0, 0}

{-1, 0, 1, 0}

{-1, 1, -1, 0}

{-1, 1, 0, 0} The variants of the invention explained so far can be combined to form a further variant in which, in addition to pulsed weight functions, still different higher-order signal forms are used. In this way, the advantages of the aforementioned variants can be further increased. Thus, different filter characteristics can be achieved with improved suppression of interference spectra. Exemplary timing diagrams of the control signal for the bias current source 42 and a resulting

Transmission characteristics of the converter stage 40 are shown in Figures 6 and 7. FIG. 6 shows, in an upper time curve, a time diagram with a pulsed triangular windowing of the bias current I _B ias, and in a lower diagram a hanning function-time curve of the bias current I _B i _as - all switching patterns of the two time profiles mentioned above are based on switching intervals I , II and III together. The advantage of using such Schaltmuster- or cycles with control signals of higher order is apparent from Figure 7, which shows a transmission behavior of the converter stage 40 with a pulsed triangular windowing of the bias current source 42. Compared to the weighting shown in FIG. 5 with a rectangular control signal waveform, a further improved suppression of broadband noise components can be achieved when weighting with a triangular window. Qualitatively, this improved suppression is apparent from a comparison of Figures 5 and 7, each with identical switching patterns.

In summary, a converter stage for detecting a sensor signal, in particular a signal of a yaw rate sensor has been presented, with which it is possible to achieve a significantly higher aliasing suppression compared to conventional systems or a broadband noise suppression within the baseband in the case of sampled and subsampled or bandpass-sampled signals. ten sensor systems. The performance increase is achieved by a specifically designed amplifier stage 41 with a bias current source 42 in combination with an electrical integration circuit, with which it is possible to apply different weight functions to the output signal lout of the amplifier stage 41 corresponding to the sensor signals and from this an improved transmission behavior of the converter stage 40 to generate. Disclosed were three different embodiments of the invention. A first variant represents a controlled amplifier stage 41 for broadband noise suppression. A further variant represents an amplifier stage pulsed with rectangular signals 41 for realizing different filter charac- istics. A third variant represents a combination of these two variants.

The invention presented here can also be used in sensors with other than capacitive mechanism of action, for example, piezoresistive rotation rate sensors, wherein both continuous-time and time-discrete implementations of the C / U converter 30 are possible.

Of course, the invention is also advantageously applicable to sensor signals derived from multi-axis angular rate sensors. As front-end solutions within an ASIC in which the invention can be applied, discrete-time as well as purely analog variations come into consideration. In light of the above explanations, it goes without saying that the embodiment of the invention is not limited to the examples shown in the figures, but is also possible in a variety of modifications, which are within the purview of the skilled person.

Claims

claims

1 . Converter stage (40) for detecting a sensor signal, in particular a signal of a rotation rate sensor, comprising an amplifier stage (41) with an electronic circuit, characterized in that different weight functions can be generated by means of the electronic circuit, with which a filter characteristic of the converter stage (40) can be influenced ,

Second converter stage (40) according to claim 1, characterized in that the electronic circuit comprises a controllable current source (42) and an electrical charge integrator (43, 44, 46), wherein the electrical charge integrator (43, 44, 46) from an output the amplifier stage (41) is controllable.

Third converter stage (40) according to claim 2, characterized in that the amplifier stage (41) is designed as a transconductance amplifier and the current source (42) is designed as a bias current source for the amplifier stage (41).

4. converter stage (40) according to any one of claims 2 or 3, characterized in that the current source (42) from a via a control line (45) transmitted, rectangular shaped control signal can be controlled.

5. converter stage (40) according to any one of claims 2 or 3, characterized in that the current source (42) of a via the control line (45) transmitted, triangular or a corresponding to a Hanning function formed control signal is controllable.

6. converter stage (40) according to claim 5, characterized in that the control signal has periodic switching cycles with N substantially equal intervals. Transducer stage (40) according to claim 6, characterized in that several of the periodic switching cycles can be combined to form differently designed switching patterns.

Converter stage (40) according to one of claims 2 or 3, characterized in that the current source (42) alternatively by a via a control line (45) transmitted rectangular, triangular, or according to a Hanning function formed control signal is controlled, wherein the control signal periodic switching cycles having N substantially equal intervals.

Transducer stage (40) according to claim 8, characterized in that several of the periodic switching cycles can be combined to form differently designed switching patterns.

Converter stage (40) according to one of claims 2 to 9, characterized in that an integration time constant of the electrical charge integrator (43, 44, 46) is variably adjustable.

Method for detecting a sensor signal, in particular a signal of a rotation rate sensor (20), the method comprising the following steps:

- detecting a capacitance of the sensor (20);

- generating a voltage dependent on the capacitance of the sensor (20);

- generating a current depending on the voltage;

- integrating an electric charge of the current over a predetermined period of time; and

- Detecting an output voltage as a sensor signal in dependence on the size of the integrated charge.