Aa current-controlled oscillator for sensor instrumentation based on a current-inversion switch

Abstract

The present application describes a current-controlled oscillator for sensor instrumentation based on a current-inversion switch (10), which, in the general embodiment, comprises a first current-source terminal input (11) and a second current-source terminal input (12) connected in parallel with a floating current-source (3) by a first terminal (3a) and a second terminal (3b), the first current-source terminal input connected to the first terminal and the second current-source terminal input connected to the second terminal; a first bias voltage input (Vb1) connected to a second electronic switch (16) and second bias voltage input (Vb2) connected to a first electronic switch (15), the first electronic switch (15) and the second electronic switch (16) being actuated by a control terminal (14), the first electronic switch (15) commuting the first current-source terminal input (11) between the second bias voltage (Vb2) and a live terminal (13), and the second electronic switch (16) commuting the second current-source terminal input (12) between the first bias voltage (Vb1) and the live terminal (13).

Description

A CURRENT-CONTROLLED OSCILLATOR FOR SENSOR INSTRUMENTATION BASED ON A CURRENT-INVERSION SWITCHTechnical Field

[0001] The present invention discloses a Current-Controlled Oscillator (CCO) for sensor instrumentation, based on a current-inversion switch.Background Art

[0002] Presently, the working principle of generic linear current-controlled oscillators can be divided in two main categories: current-steering oscillators (FIG. 1 ), and charge-balance oscillators (FIG. 2). Usually, the current-steering principle is simpler than the charge-balance one, being easier to implement.

[0003] Generic current-steering CCOs are usually implemented in CMOS technology, where an input current I is mirrored (and often amplified) through the P-MOS and N-MOS mirrors for making the charge and discharge currents, which are sequentially applied to the integrating capacitor C by the alternate action of the switches. As these currents are usually symmetrical, the capacitor voltage Vi is triangular, having an amplitude AV = Vih - Vil, and a frequency fo. The accuracy and linearity of this generic current-steering CCO implementation depend mainly on the accuracy of the current-mirrors over the desired dynamic-range and temperature. As so, mirror design is critical in wide dynamic-range applications. The operation at very-low current levels in the generic current-steering CCO is particularly difficult, because some MOSFETs must work in the subthreshold regime, which is very sensitive to process and temperature variations, severely compromising the performance of the mirrors. Since photodiode currents can go down to picoampere levels in very-low light applications, this approach is, therefore, not recommended for such applications.

[0004] In the particular application of photodiode instrumentation, and due to the above-mentioned technical limitations, the charge-balance principle is often used. In the charge-balance CCO, in the steady state, the voltage at the capacitor ( Vi) is a sawtooth with an amplitude A V During the integration time Ti, the switch is open, and the photodiode current / is integrated in the capacitor. When Vi reaches Vp, apulse of a fixed de-integration time Td is generated by a precision monostable multivibrator, closing the switch. During this de-integration phase, the current in the capacitor is reversed, being now given by Iref- I. After Td, a new integration cycle begins, and the whole process is repeated. The relation between the current / and the output frequency fo is given by / = Iref ■ Td ■ fo. Since the parameters Iref and Td are constants set by design, this type of CCO is linear. Regarding linearity and accuracy, its performance is thus determined by the behavior of the parameters Iref and Td. However, the accurate generation of Tdis particularly difficult, requiring a precision monostable multivibrator, which is a complex block to design. Moreover, AV is given by AV = Iref - I) ■ Td / C, revealing that the amplitude of Vi depends on I, which may further difficult the design process. Finally, for a given dynamic range, the optimization of the speed requirements of the integrating operational amplifier (slew-rate and Gain - Bandwidth Product (GBW)) is not often easy to achieve, due to the asymmetrical nature of Vi slopes and parameter interdependence, demanding for a careful dimensioning of the parameters Iref, Vp, Td, and the integrating capacitor C.Summary of Invention

[0005] The present application describes a current-controlled oscillator for sensor instrumentation based on a current-inversion switch, which, in the general embodiment, comprises a first current-source terminal input and a second currentsource terminal input connected in parallel with a floating current-source which has a first terminal and a second terminal, the first current-source terminal input connected to the first terminal and the second current-source terminal input connected to the second terminal; and a first bias voltage input connected to a second electronic switch and second bias voltage input connected to a first electronic switch, the first electronic switch and the second electronic switch being actuated by a control terminal, the first electronic switch commuting the first current-source terminal input between the second bias voltage and a live terminal, and the second electronic switch commuting the second current-source terminal input between the first bias voltage and the live terminal; characterized by an unidirectional input current supplied by the floating current source being transformed into a symmetrical bidirectional output current in the live terminal.

[0006] In the general embodiment of present invention, the first electronic switch and / or the second electronic switch comprise a Double-Pole Double-Throw (DPDT) switch, or two coupled Single-Pole Double-Throw (SPDT) switches.

[0007] Yet in the general embodiment of present invention, the live terminal of the current-inversion switch is connected to a first terminal of a current integrator, wherein the output voltage of said integrator, provided in a second terminal, is characterized by comprising a perfect triangle waveform with equal integration and de-integration times.

[0008] Yet in the general embodiment of present invention, the output voltage of the integrator provided in the second terminal is connected to an input terminal of a window comparator with memory.

[0009] Yet in the general embodiment of present invention, the control terminal of the current-inversion switch is connected and driven by an output terminal of the window comparator with memory.

[0010] Yet in the general embodiment of present invention, the floating current-source comprises an arbitrary bias voltage, defined by the voltages set at the first bias voltage input, at the second bias voltage input, and at the reference voltage input of the integrator.

[0011] In another general embodiment of present invention, an output of the window comparator with memory is connected to a frequency-doubling arrangement.

[0012] In another general embodiment of present invention, the floating current-source comprises a photodiode or a similar device.

[0013] In another general embodiment of present invention, the floating current-source is replaced by a resistive element or a similar device.

[0014] In the preferred embodiment of present invention, the voltages set at the first bias voltage input, at the second bias voltage input, and at the reference voltage input of the integrator are equal, and the bias voltage of the floating current-source is zero, enabling the first electronic switch and the second electronic switch to operate in a no-voltage condition when in the off-state, thus minimizing switch leakage-currents.

[0015] The present application also describes a light-to-frequency converter, characterized by comprising the current-inversion switch previously described.

[0016] The present application also describes a current-controlled oscillator for sensor instrumentation, characterized by comprising the current-inversion switch previously described.Technical Problem

[0017] Based on prior art limitations, it becomes clear that the generic current-steering CCO is not appropriate for very-low current applications. On the other side, the charge-balance oscillator principle requires the generation of a very precise T in the de-integration cycle, involving a considerable complexity.Solution to Problem

[0018] To overcome the identified problem, and also the limitations of prior art, it is proposed the direct use of the photodiode current both for the integration and deintegration cycles, resulting effectively in a current-steering topology. This will be achieved by reversing the input current of the integrator with an electronic switch, as depicted in FIG. 3.Advantageous Effects of Invention

[0019] Based on the above, due to the proposed current-inversion switch, the magnitude of the charging and discharging currents is the same, and the waveform of the output voltage of the integrator is now a perfect triangle with integration and de-integration times being equal, Td = Ti, as characteristic of a current-steering topology. However, in contrast to the typical current-steering topology of FIG. 1 , there is no mirroring or amplification of the photodiode current since this current is now fed directly to the integrator. This way, the errors associated with mirroring and amplification are completely eliminated, which is a definite advantage in very- low photodiode current applications, in the range of picoampere level, which may occur in very-low light applications or environments, recalling that such errors can be considerable in these applications, as already exposed.

[0020] The relation between the current I and the output frequency fo is given by I = C ■ AK ■ 2 ■ fo, being C the capacitance of the integrating capacitor, and Al / the amplitude of the triangular waveform at the output of the integrator, which is givenby AV = VH - VL. Since the parameters C and A V are constants set by design, this CCO is linear.

[0021] In the proposed development, and in respect to linearity and accuracy, the performance of the physical circuit implementation is mainly determined by the behavior of parameters C and A V. As A V = VH - VL, its accuracy will depend on VH and VL, which are usually derived from some reference voltage Vref. Comparing the invention with the charge-balance approach, the complexity involved on obtaining an accurate Vref and Iref is similar; however, it is usually simpler to implement an accurate capacitor than generating a precision timing for Td. The speed requirements of the integrating operational amplifier are also simpler to determine, as the slopes of the output voltage of the integrator just depend on / and the integrating capacitor C, being symmetrical.

[0022] The current-inversion switch of the proposed CCO is, therefore, the core element of the invention, being used for transforming the unidirectional input current I, provided by a photodiode or similar device, into a bidirectional current (+ / ) that feeds the integrator. In its simplest form, the proposed current-inversion switch only has two states. Internally, the switch comprises a double-pole double-throw (DPDT) type switch, or two coupled single-pole double-throw (SPDT) type switches, which change state synchronously.Brief Description of Drawings

[0023] For better understanding the present application, figures representing the most relevant embodiments are herein attached which, however, are not intended to limit the technique disclosed herein.

[0024] The current-controlled oscillator for sensor instrumentation based on a currentinversion switch will now be described in greater detail based on non-limiting exemplary embodiments and with reference to the figures, on which:

[0025] [FIG. 1 ] discloses an illustration of the background art, an example of the current-steering method of a CCO.

[0026] [FIG. 2] discloses an illustration of the background art, an example of the charge-balance method of a CCO.

[0027] [FIG. 3] discloses an illustration of the general embodiment of the invention, in the first operational state.

[0028] [FIG. 4] discloses an illustration of the general embodiment of the invention, in the second operational state.

[0029] [FIG. 5] discloses an example of a simple frequency-doubler based on a XOR logic-gate.

[0030] [FIG. 6] discloses an example of the application of bootstrapping techniques to a SPST switch for reducing its leakage-current, resulting in a T-structure switching arrangement.

[0031] [FIG. 7] illustrates another embodiment of the current-inversion switch (10) resorting to the use of T-structures for reducing switch leakage-current, in the first and second operational states.

[0032] [FIG. 8] discloses the preferred embodiment of the invention, which arises from the particular case where Vb = 0, Vb1 = Vi = Vb2 = 0. This is the simplest embodiment of the invention.

[0033] [FIG. 9] details the implementation of the preferred embodiment with elementary SPST switches, when in the first operational state.

[0034] [FIG. 10] details the implementation of the preferred embodiment with elementary SPST switches, when in the second operational state.

[0035] [FIG. 1 1 ] details the implementation of the preferred embodiment with elementary SPST switches, when in an optional reset state.Description of Embodiments

[0036] With reference to the figures, some embodiments are now described in more detail, which are however not intended to limit the scope of the present application.

[0037] The FIG. 3 and FIG. 4 disclose the general embodiment of the invention, where it is possible to identify three major blocks: the current-inversion switch (10), a current integrator (20), and a window comparator with memory (30).

[0038] In the embodiment of FIG. 3 and FIG. 4, the current-inversion switch (10) comprises a first current-source terminal (11 ), a second current-source terminal (12), a live terminal (13), a control terminal (14), a first electronic switch (15) and asecond electronic switch (16). These switches can be viewed as part of a single DPDT switch, or as two coupled SPDT switches. Additional inputs are included for bias voltages, particularly identified as first bias voltage ( Vb1) and second bias voltage ( Vb2). The integrator (20) comprises a first terminal (21 ), a second terminal (22), an operational amplifier (23), a capacitor (24), an optional reset switching arrangement (25), and a reference voltage ( Vi) input. The integrator (20) is a generic current integrator, where the current to be integrated is supplied at the first terminal (21 ), which is a low-impedance node fixed at a reference voltage ( Vi), applied at the Vi input terminal. The current integrator (20) implementation shown is the most common one, and it should be considered just as a typical implementation example. In fact, any implementation of a current integrator (20) with the aforementioned characteristics can be used instead. The window comparator with memory (30) comprises a first reference voltage ( VH) input, a second reference voltage ( VL) input, an input terminal (31 ), a first comparator (33), a second comparator (34), a SR flip-flop (35) with an output terminal (32). The implementation of the window comparator with memory (30) shown is just for describing its functionality; alternatively, any implementation with a similar functionality can be used. As illustrated, an optional frequency-doubling arrangement (50), with an input and output (51 ) can be serially connected with the comparator (30) to double the frequency of the oscillator. Finally, a photodiode current-source (3) comprises a cathode terminal (3a) and an anode terminal (3b), which are connected to the first current-source terminal (1 1 ) and to the second current-source terminal (12) of the current-inversion switch (10), respectively. It should be noted that any general floating current-source can used instead of the photodiode, broadening the application scope of the invention.

[0039] The oscillator has at least two possible states, set by the control signal U / D, which is generated by the flip-flop (35). These states are depicted in FIG. 3 and FIG. 4, noting that these figures just differ on the state of the current-inversion switch (10), according to the state of the signal U / D. In the following description, it is assumed that the photodiode (3) current I is constant during an oscillation period, and that the reset switch (25) of the integrator is open.

[0040] In the first state, ramp up (U / D = 1 ), illustrated in FIG. 3, the cathode of the photodiode (3a) is connected to the first terminal (21 ) of the integrator, and theanode of the photodiode (3b) is connected to the bias voltage Vb1. As so, the photodiode (3) current I comes out of the first terminal (21 ) of the integrator, charging the capacitor (24), and the voltage at the second terminal (22) increases linearly. When the voltage at terminal (22) rises above the first reference voltage of the comparator ( VH), the comparator (33) changes the state of the flip-flop (35), and the oscillator enters the second state, ramp down (U / D = 0).

[0041] In this second state, illustrated in FIG. 4, the cathode of the photodiode (3a) is connected to the bias voltage Vb2, and the anode of the photodiode (3b) is connected to the first terminal (21 ) of the integrator. As so, due to the currentinversion action of the switch arrangement (10), the photodiode current I now enters the first terminal (21 ) of the integrator, discharging the capacitor (24), and the voltage at the second terminal (22) decreases linearly. When the voltage at terminal (22) falls below the second reference voltage of the comparator (VL), the comparator (34) changes the state of the flip-flop (35) again, and the oscillator returns to the first state, ramp up (U / D = 1 ), completing an oscillation period.

[0042] In some embodiments, it can be desirable to have a way of stopping the oscillator, or starting it from a known state. For achieving this, an optional reset state can be implemented, besides the two aforementioned states. This can be accomplished by the reset switch (25) of the integrator, for example. In some other embodiments, it can be desirable to have the photodiode (3), or equivalent floating current-source, under certain operating conditions during the reset state. In this case, a customized reset state could also be implemented in the current-inversion switch (10), according to the state of the RST signal (the control signal of the switch (25)).

[0043] As previously exposed, the photodiode (3) current is directly used for charging and discharging the integrating capacitor (24), leading to a perfect triangular waveform (with symmetrical slopes) at the second terminal (22) of the integrator, having a frequency fo. This is the oscillation frequency of the oscillator, being also the frequency at the output of the flip-flop (35). The relation between the photodiode current / and the output frequency fo is given by I = C ■ AK ■ 2 ■ fo, being C the capacitance of the integrating capacitor (24), and AV the amplitude of the triangular waveform at the second terminal (22) of the integrator, which is given by AV = VH - VL

[0044] The perfect triangular waveform at the second terminal (22) of the integrator enables another improvement. It may be advantageous to count ramps instead of full cycles at the oscillator output, as the number of occurrences per time unit is bigger, therefore increasing the frequency resolution. In this case, the output frequency fo is given by I = C ■ AV ■ fo, being now fo defined as the number of ramps per time unit. However, ramp counting requires dual-edge counting at the oscillator output, which involves special circuitry. Alternatively, a frequency-doubler (50) can be inserted at the output of the oscillator, enabling the normal single-edge counting at the output (51 ). An example of a simple frequency-doubler circuit based on a XOR logic gate is shown at FIG. 5.

[0045] The general embodiment of the invention supports the operation of the photodiode (3) at an arbitrary bias voltage Vb. The polarity of this voltage is defined as usual: Vb is positive if the anode (3b) of the photodiode (3) is positive relatively to its cathode (3a); otherwise, Vb is negative. Similarly, the photodiode (3) current I is positive if it enters the photodiode anode (3b); otherwise, it is negative. The photodiode (3) bias voltage Vb is set by the aforementioned bias voltages Vb1, Vb2, and Vi, as follows:• Vb < Q, Vb1 < Vi < Vb2• Vb = Q, Vb1 = Vi = Vb2• Vb > Q, Vb1 > Vi > Vb2

[0046] Moreover, the following relations must apply:• Vb = (Vb1 - Vb2) / 2• Vi = (Vb1 + Vb2) / 2.

[0047] In FIG. 3 and FIG. 4, the direction of the photodiode (3) current / is represented as being negative. This is the most usual case, as in general the photodiode is operated at Vb < 0. If the photodiode (3) is operated at a positive Vb, the direction of the photodiode (3) current will be reversed, and the control signal (14) of the current-inversion switch (10) should be inverted accordingly, otherwise the oscillator will not work. In other words, if Vb > 0, the control signal U / D should become U / D, by deriving it from the inverted output of the flip-flop (35), for example.

[0048] The switching arrangement of the current-inversion switch (10) and the optional reset switch (25) will be ultimately implemented with elementary single-pole singlethrow (SPST) type switches. Depending on the bias of the photodiode ( Vb), some of these SPST electronic switches can be under a considerable voltage when in the open-state. This voltage may originate some leakage current through the switch, which may affect the linearity of the oscillator in applications that involve a very-low photodiode current. In such cases, the switch leakage current can be minimized by nulling the voltage across the switch in the open-state, which can be done by using bootstrapping techniques. A classic way of implementing such techniques on a SPST switch is shown in FIG. 6, in the simpler case of a switch that only has a sensitive node (SN) and a robust node (RN). The SPST switch is replaced by an arrangement that has a unity-gain voltage-buffer with high inputimpedance, and three SPST switches that make a T-structure. A sensitive node is considered to be a node that deals with high-impedance and / or low-level signals, being in general an input node, whereas a robust node is a low-impedance power node (a voltage source or a supply), being usually an output node. When the switch is closed (FIG. 6 a)), the output of the buffer is disconnected, and the buffer plays no role in the circuit. When the switch is opened (FIG. 6 b)), the voltage across the SPST switch of the sensitive node is nulled by the buffer, thus minimizing the leakage current in that switch; meanwhile, the leakage current in the SPST switch of the robust node is diverted to the buffer.

[0049] The application of the T-structure of FIG. 6 to the current-inversion switch (10) of FIG. 3 and FIG. 4 yields another embodiment of the current-inversion switch. This embodiment is shown in FIG. 7, for the ramp-up state (FIG. 7 a)), and the ramp-down state (FIG. 7 b)). The current-inversion switch (10) is now implemented by an arrangement of four T-structures. The sensitive nodes are the photodiode terminals (3a) and (3b), and the live terminal (13) of the current-inversion switch (10). The robust nodes are the bias voltages Vb1, Vb2, and Vi. Due to the fact that the live terminal (13) of the current-inversion switch (10) is connected to the first terminal (21 ) of the integrator (20), which is at a fixed potential Vi, the T-structure arrangement can be made in a way that dismisses all the bootstrapping buffers, considerably simplifying the resulting circuit.

[0050] The different possibilities regarding the bias voltage Vb of the photodiode enable a myriad of embodiments, which are particular cases of the general embodiment of FIG. 3 and FIG. 4. Amongst them, the case of Vb = 0 with Vb1 = Vi = Vb2 = 0 is perhaps the most interesting one, as it enables a quite simple implementation that has a high-performance at very-low photodiode currents, in the picoampere range. This embodiment of the invention ( Vb = 0, Vb1 = Vi = Vb2 = 0) is the preferred embodiment, being represented in FIG. 8. Besides dismissing all the circuitry necessary for the generation of the bias voltages Vb1, Vb2, and Vi, all the elementary switches of the current-inversion switch (10) operate at no voltage in the off-state, meaning that no T-structures are needed for minimizing switch leakage-currents. Moreover, as the common-mode voltages on the currentinversion switch are at ground level, or very-close to it, each of the elementary SPST switches of (10) can be implemented with a simple N-channel MOSFET, further simplifying the resulting circuit.

[0051] The detailed action of the current-inversion switch (10) together with the reset switching arrangement (25) for the preferred embodiment is presented in FIG. 9, FIG. 10, and FIG. 1 1 , revealing the implementation with elementary SPST electronic switches. In the current-inversion switch (10), the SPDT switch (15) of the general embodiment is implemented by the elementary SPST switches (15a) and (15b); the SPDT switch (16) of the general embodiment is implemented by the elementary SPST switches (16a) and (16b). The reset switching arrangement comprises a T-structure for minimizing the aforementioned leakage-currents, being implemented by the elementary SPST switches (250), (251 ), and (252). The FIG. 9 shows the first state, ramp up (U / D = 1 ), whereas FIG. 10 shows the second state, ramp down (U / D = 0). The action of first and second states is identical to the case of the general embodiment, which was already extensively described. The FIG. 11 details the optional reset state (RST = 0). This third state can have several implementations, as previously stated. In the embodiment of FIG. 1 1 , the currentinversion switch (10) is configured to disconnect the photodiode (3) from the integrator, by opening the elementary switches (15b) and (16b), to avoid the transient that occurs when the reset switch arrangement (25) closes. Additionally, during the reset state, the photodiode (3) is kept at null bias ( Vb = 0) and grounded, by closing the elementary switches (15a) and (16a), to keep the working conditionsthe photodiode has during normal operation (oscillation). The objective is to maintain the photodiode as undisturbed as possible when changing from an operational state (oscillation, first or second state) to the reset state, and vice-versa.Examples

[0052] The invention is particularly useful and advantageous in very-low current measurements, which can go down to the picoampere level, enabling a very-wide linear dynamic-range. The main application field of the invention is photodiode or phototransistor instrumentation; however, it is extensible to any device that can work as a floating current-source. Moreover, due to the ability to impose a fixed operating bias voltage ( Vb), the invention can work also with floating resistive devices, where it translates a resistance into a current, and then into an oscillation frequency. In short, the invented oscillator can convert current or resistance from a floating device into frequency, enabling the instrumentation of a wide range of sensors. In addition to optical measurements (photometry, spectrometry), the application list includes temperature measurement, deformation and / or strain measurement, weighting systems, pressure sensors, instrumentation of chemical sensors (optical or resistive).Industrial Applicability

[0053] The invention is particularly suited to be integrated on-chip, in CMOS technology, thus targeting the semiconductor industry. An implementation with discrete components is also possible, but with some performance degradation on the achievable speed.

Claims

Claims

1. Current-inversion switch (10) comprising a first current-source terminal (1 1 ) input and a second current-source terminal (12) input connected in parallel with a floating current-source (3) by a first terminal (3a) and a second terminal (3b), the first current-source terminal (1 1 ) input connected to the first terminal (3a) and the second current-source terminal (12) input connected to the second terminal (3b); and a first bias voltage (Vb1 ) input connected to a second electronic switch (16) and second bias voltage (Vb2) input connected to a first electronic switch (15), the first electronic switch (15) and the second electronic switch (16) being actuated by a control terminal (14), the first electronic switch (15) commuting the first current-source terminal (1 1 ) input between the second bias voltage (Vb2) and a live terminal (13), and the second electronic switch (16) commuting the second current-source terminal(12) input between the first bias voltage (Vb1 ) and the live terminal (13); characterized by an unidirectional input current supplied by the floating current-source (3) being transformed into a symmetrical bidirectional output current in the live terminal(13).

2. Current-inversion switch (10) according to previous claim 1 , characterized by the first electronic switch (15) and / or the second electronic switch (16) comprising a Double-Pole Double-Throw switch, or two coupled Single-Pole Double-Throw switches.

3. Current-inversion switch (10) according to previous claim 1 , wherein the live terminal (13) is connected to a first terminal (21 ) of a current integrator (20), wherein the output voltage of said integrator (20), provided in a second terminal (22), is characterized by comprising a perfect triangle waveform with equal integration and de-integration times.

4. Current-inversion switch (10) according to previous claim 3, wherein the output voltage of the integrator (20), provided in the second terminal (22), is connected to an input terminal (31 ) of a window comparator with memory (30).

5. Current-inversion switch (10) according to previous claim 1 and 4, wherein the control terminal (14) is connected and driven by an output terminal (32) of the window comparator with memory (30).

6. Current-inversion switch (10) according to previous claim 1 , wherein the floating current source (3) comprises an arbitrary bias voltage (Vb), defined by the voltages set at the first bias voltage (Vb1 ) input, at the second bias voltage (Vb2) input, and at the reference voltage (Vi) input of the integrator (20).

7. Current-inversion switch (10) according to previous claim 6, wherein the voltages set at the first bias voltage (Vb1 ) input, at the second bias voltage (Vb2) input, and at the reference voltage (Vi) input of the integrator (20) are equal, and the bias voltage (Vb) of the floating current-source (3) is zero, enabling the first electronic switch (15) and the second electronic switch (16) to operate in a no-voltage condition when in the off-state, thus minimizing switch leakage-currents.

8. Current-inversion switch (10) according to previous claim 5, wherein an output of the window comparator with memory (30) is connected to a frequencydoubling arrangement (50).

9. Current-inversion switch (10) according to previous claim 1 , wherein the floating current-source (3) comprises a photodiode or a similar device.

10. Current-inversion switch (10) according to previous claim 1 , wherein the floating current-source (3) is replaced by a resistive element or similar device.[Claim 1 1 ] Light-to-frequency converter, characterized by comprising the currentinversion switch (10) described in any of the previous claims 1 to 10.

12. Current-controlled oscillator for sensor instrumentation, characterized by comprising the current-inversion switch (10) described in any of the previous claims 1

Images