WO2010115497A1

WO2010115497A1 - Sensor readout circuit, sensor and method for reading out a sensor element

Info

Publication number: WO2010115497A1
Application number: PCT/EP2010/001466
Authority: WO
Inventors: Hartmut Jesch; Raimund Barden; Dirk Steigel; Willi Reichert
Original assignee: Perkinelmer Optoelectronics Gmbh & Co. Kg
Priority date: 2009-03-30
Filing date: 2010-03-09
Publication date: 2010-10-14
Also published as: DE102009015586A1

Abstract

A sensor readout circuit (2) comprises an input terminal (11) for a sensor element (1) for electromagnetic radiation producing an electric sensor signal, an integrator (12) for integrating the sensor signal, a reset means (13) for resetting the integrator, and a control circuit (14) for repeatedly operating the reset means after an integrating period. A buffer capacitor (15) is connected to the input terminal. A buffer switch (16) is operable by the control circuit for connecting the input terminal with the integrator input within the integration period and disconnecting it therefrom during integrator reset.

Description

Sensor Readout Circuit, Sensor and Method for Reading out a Sensor Element

The invention relates to a sensor readout circuit, a sensor and a method for reading out a sensor element according to the preamble of the independent claims.

Relevant prior art is the Hamamatsu data sheet "Photosensor Modules H 7468 Series". Figures 4a and 4b of the present application show content of the Hamamatsu data sheet. Basically, as schematically shown in Fig. 4a, the overall sensor has a photomultiplier tube 1, a readout circuit 2 and a driving circuit 3. The components are interconnected. The readout circuit 2 receives the electric signal from the photomultiplier tube. The photomultiplier tube is driven by the driving circuit. Arrow 4 represents sensed incident radiation.

Fig. 4b shows the components in detail: 1 is the photomultiplier tube with a photosensitive cathode Ia, an anode Ib and a dynode chain Ic along which a voltage drop occurs for accelerating electrons emitted from the cathode Ia and multiplying them towards the anode Ib where they appear as individual avalanches for protons incident on the cathode. The connection scheme is as shown in Fig. 4b. A multistage voltage rectifier circuit is provided, an oscillator circuit and a stabilizing circuit which receives high voltage feedback and has a voltage control input. The multiplier output (i.e. the anode terminal) is connected to an integrator input so that the signal is integrated. The integrator has an operation amplifier and a capacitor in the feedback branch. A short circuiting switch is connected in parallel to the capacitor. In usual manner, the integrator integrates the signal from the photomultiplier tube (PMT) , and the integrator is repeatedly/periodically reset by closing (making conductive) the switch. A microcontroller is provided that receives the A/D converted output signal of the integrator and operates the switch and also the driving circuit. The minimum dead time of the known module is 10 μs (microseconds) . The minimum integration time is 40 μs . The dead time is required for sampling the output signal and for resetting the integrator.

The disadvantage of this known approach is that its measurement accuracy is systematically limited. In the dead time, the PMT output signal is systematically lost and thus not measured. This error becomes the more important the more the dead time is comparable to the integration time because then the share of "lost signal" becomes relatively large.

A solution would be to make relatively long integration times compared to the reset time so that the relative amount of "lost signal" decreases. Further, the longer the integration time is, the worse is the resolution in time. But long integration times constitute a problem for high sensor element output signals (i.e. high integrator input signals) . The i i i L c y j- α t v-' j. u nci i α oπc o

o α t u± α i αnu non-linear so that its output is no longer representative of the sensed input quantity.

When running into saturation, the sensor can be readjusted through the driving circuit of the sensor element so that the output signal of the PMT decreases for the same input quantity 4 so that saturation of the integrator is avoided. Then, however, also the driving conditions of the driving circuit 3 must be considered for evaluating the integrator output. And besides, adjusting the driving circuit 3 may require time during which the output signal can not be used and may even be too slow for following quickly changing PMT output signals.

Other known prior art is CN1221812, CN1621812A, FR2848676A1, GB2350187A, JP2001004445, JP2004125626, JP57044821, JP 63201587, US2004169128A1, US2006202126A1, US5773816A1, US5920199A1, US6384401B1, US6501322B1, US6642501B2, US7339153B2, USRE39527E, WO2006037248A1 and US6166365.

It is the object of the present invention to provide a sensor readout circuit, a sensor and a readout method having improved accuracy across a wide range of input signal intensity. This object is accomplished by the features of the independent claims. The dependent claims are direct on preferred embodiments of the invention.

According to the invention, a readout circuit for a sensor element has a buffer capacitor at the input side of the readout circuit. The sensor element produces an electric signal collectable as electric charge. The buffer capacitor receives the signal from the sensor element and stores it during dead time of the readout circuit, within which the buffer capacitor is disconnected from the ^Λλdownstream" readout circuit. After the dead time, the buffer capacitor is re- connected to the readout circuit, and its signal is evaluated there. Through this measure, the influence of the dead time is significantly reduced because signal loss during dead time is significantly reduced, if not made zero. Accordingly, it is possible to make cycle times of integration comparatively short without encountering the problem of increased significance of signal loss during dead time because signal loss during dead time is suppressed. Due to the shorter integration cycle, also high output signals of the sensor element can be measured without operating the settings so that the above described disadvantage is avoided.

A sensor has a sensor element and a readout circuit as mentioned above. In such a sensor, the integration cycle (overall steady state period) may be lower than 30, 20, 15, 10 or 5 μs . It may be higher than 100 ns, 200 ns, 500 ns, 1, 2 or 5 microsecond. A preferred range is presently 8 to 12 μs, but may also be one of the above mentioned values +/- 20% thereof.

The sensor may be a sensor for electromagnetic radiation or for particles (which may be converted into radiation by an appropriate input window material) .

In the following, embodiments of the invention will be described with reference to the attached drawings, in which

Fig. 1 is a basic circuit of a readout circuit according to the invention,

Fig. 2 shows driving patterns in the method of reading out a sensor element, Fig. 3 shows characteristics of the overall sensor, and

Fig. 4 relates to prior art.

Generally, in this specification same reference numerals shall denote same features. Features in this specification shall be deemed combinable with each other even if this is not explicitly said, as far as a combination is not excluded for technical reasons. Apparatus features shall also be considered as a disclosure for a method step implemented by said apparatus feature, and vice versa, a method aspect shall also be deemed to refer to means for implementing said method aspect.

In Fig. 1, the readout circuit 2 has an input terminal 11 to which a sensor is connected or connectable. The terminal 11 may explicitly be provided or may be an integral part of some kind of circuitry. The sensor element 1 may be a photomultiplier tube, sensitive to radiation or particles and producing electron avalanches upon incidence of radiation or particles, symbolized by arrow 4. The photomultiplier may be built as described in US6166365.

Close to the input side of the readout circuit 2, and preferably connected to its input terminal 11, is a buffer capacitor 15, which may, with its other terminal, be connected to a fixed potential, for example ground. A buffer switch 16 is provided for selectively connecting the input side/input terminal 11, and thus also the buffer capacitor 15, to the "downstream" part of the readout circuit 2. It comprises an integrator 12 which may be built by an analogue amplifier 17 and a feedback capacitor 18, and an analogue signal shaping/evaluation circuit 14 "downstream" of the integrator 12, e.g. for filtering, amplitude matching. The amplifier 17 may be an operational amplifier with an inverting input and a non-inverting input. The non-inverting input may be connected to a fixed potential, for example again ground.

The inverting input receives the feedback signal from the feedback capacitor 18, and is also connected to the buffer switch 16 so that it may receive the signal from the sensor or buffer capacitor 15 when buffer switch 16 is closed (conductive) . The output of the operational amplifier 17 is connected to the feedback capacitor 18 and is further connected to a suitable signal shaping/evaluation circuit 14 which may also have control purposes for the circuit, particularly for the readout circuit, but potentially also for the sensor driving circuit 3 (not shown in detail). The evaluation circuit 14 may comprise a sample-and-hold circuit and may comprise an analogue circuit. But likewise, it may comprise an analogue/digital converter 10 at the output side of the integrator 12 and may comprise digital components 19a such as digital sampling in a sampling circuit, and digital control in a control circuit 19b.

In between the output of the integrator and the input of the analogue/digital converter there may be some kind of analogue circuitry e.g. for adapting the output signal of the integrator to the requirements of the input of the analogue/digital converter. Further, a reset means 13 is provided which may be a reset switch operable under control of the control circuit 19b. The switch may be in parallel to capacitor 18, i.e. between output and inverting input of amplifier 17. Once closed (conductive), the reset switch 13 short circuits capacitor 18 and thus brings the output of the integrator 17 quickly to the potential of the non-inverting input, which may be the fixed potential, e. g. ground, as said above.

A very fundamental operation scheme of the above circuit may be that during reset (reset switch 13 closed (conductive)) buffer switch 16 is open (non- conductive) , so that the electric signal from the sensor element goes into the buffer capacitor 15, but is not short circuited and is thus not lost for subsequent evaluation/measurement. When reset is finished, i.e. reset switch 13 is open again (non- conductive) , buffer switch 16 may be closed (made conductive) , so that the meanwhile accumulated charge in capacitor 15 is transferred to the input side of the integrator (comprising operation amplifier 17) and integrated there.

Since the input voltage across the input terminals of an operation amplifier 17 is always almost 0, the inverting input has almost the same potential as the non-inverting input. Thus, the voltage at the buffer capacitor will be drawn to more or less the voltage as the non-inverting input, which may be the same as that on the other terminal of the capacitor 15 so that the voltage across capacitor 15 may finally be almost zero so that it is discharged. But its charge has properly been integrated by integrator 17 which shows then a respective signal at its output.

Eventually, buffer switch 16 is opened again (made non-conductive) , so that charge from sensor element 1 accumulates only in capacitor 15, but is not relayed to integrator 12 which then remains at a constant value so that one measurement cycle is finished. The "downstream rest" of the readout circuit is free for organizing the other required operations such as sampling, holding, resetting (by shorting the feedback capacitor) , converting, determining control parameters, and so on.

Digital component 19a may be adapted (e. g. programmed) to performing- and/or decide-tϊw^ various evaluation tasks to be described later in detail, such as integrator output conversion, noise determination and elimination, addition of measurement cycle results, threshold checking, and higher level tasks such as monitoring, organizing and executing calibration, mode switching, measurement cycle input and decision, input/output organisation, internal administration and timing, interfacing towards higher level controls. Component 19b may be provided for controlling the readout circuit (particularly implementing switching schemes) and for controlling the sensor element drive circuit 3. It may have digital and/or analog outputs and may be in communication with digital circuit 19a.

Fig. 2 shows macroscopic and detailed operation schemes of the readout circuit. Macroscopically, the driving scheme may comprise an initiation cycle 21, one or more variable length measurement cycles 22, and a sequence of fixed length measurement cycles 23a, 23b and following.

The diagram shows at line 24 a fundamental operation control signal which may be supplied from external, at 25 the operation of buffer switch 16, at 26 the operation of reset switch 13 and at 27 trigger signals. In the diagrams of fig. 2, a high level in lines 25 and 26 indicates a closed switch (conductive), whereas a low level indicates an open switch (non- conductive) .

The initiation cycle 21 is a series of openings and closings of the switches 13 and 16 for erasing possibly accumulated charge. The purpose is to bring the circuit to a defined state before measurement begins .

Cycle 22 is a variable length measurement cycle which may be, in its basic operation scheme, the same as the subsequent fixed length measurement cycles, but has a variable length and thus different durations of several of its individual steps. With this, overall measurement times can be freely set, as follows: If a fixed cycle length (of measurement cycle 23a and following) is, for example, 10 μs, with this feature alone measurement times of n * 10 μs (i.e. 10, 20, 30, 40, ... μs) could be implemented. It is noted here that due to the integration the measurement results of individual measurement cycles 22, 23 can be added in a downstream stage of signal evaluation for rendering a measurement signal equivalent to that of one large measurement period composed by plural fixed or variable length cycles. If, e. g., the fixed length cycles have 10 microsecond duration, the measurement period can only be adjusted in ten microsecond steps. Here, the variable length cycle 22 helps to obtain a finer scaling. The variable length cycle 22 may, for example, be variable in its length (e. g. in dependence of an external control signal) between 10 μs and 20 μs in, say, 0,1 microsecond steps.

It is noted in this context that the time constants encountered presently are such that it is hard to implement overall measurement cycles below a certain value (presently in the range of some μs) for rendering the. desired short integration time while still being able to accomplish in such a comparatively short duration also the organizational steps (sampling, reset, drive settings ... ) necessary for properly operating the overall sensor. But it is well possible to increase an individual cycle by values much smaller than the minimum cycle duration. So, although a technical minimum cycle duration may be, for example, some few (e. g. 3) μs, and a fixed length cycle duration may be set to 5 or 10 μs, it is well possible to increase the cycle duration above said technical minimum value in steps much smaller than the technical minimum cycle duration, for example in steps of 0,1 μs

So, with the variable length cycle 22, a user is able to fine-tune the overall measurement duration according to his requirements. Instead of only the first cycle being variable in length, all of them may be variable and may be set in duration in accordance with an input signal. If for example for some reasons a measurement duration of 27 μs is required, the user may make his layout such that for the variable length cycle a duration of 7 μs is adjusted which may be taken together with the value from two fixed length cycle of 10 μs each. If an overall measurement duration of, for example, 41 microsecond is needed, the user may set the duration of the variable length cycle 22 to 11 μs and may further add to the value obtained from this cycle the values obtained from the following three fixed length cycles 23 of 10 μs each so that an overall measurement duration of 41 μs is obtained.

The overall operating method of reading out the sensor element may comprise the step of deciding the length of the measurement duration, defining the length and number of a variable length measurement cycle and the number of fixed length measurement cycles, and may further comprise the step of adding the measurement result of the defined cycles, possibly the variable length cycle amongst them. The procedure may start in accordance with a start signal, preferably from external, such as the falling edge shown in line 24.

For a significant part of the measurement cycle the input section of the readout circuit including the buffer capacitor 15 may be connected to the rest of the circuit, i.e. buffer switch 16 being closed (conductive) . In this period, the earlier accumulated signal in the buffer capacitor 15 and the sensor signal are relayed to the integrator and integrated there. A disconnection period, implemented by buffer switch 16, may be used for administration.

The details of the timings in Fig. 2 are as follows :

At least over a part of the initiation cycle 21 tl, the switches 13 and 16 may switch in opposing manner (states t5, t2, t3, t6, e.g. of 1 to 2 μs each), i.e. the one being open while the other is closed. A state where both switches are closed (conductive) may be avoided by providing time spacings t4, t7, t8 (of e.g. 50 ns to 100 ns (nano seconds) each) . This sequence is for finally bringing the integrator 12 and the buffer capacitor 15 to more or less 0.

The measurement cycle may begin with buffer switch 16 being opened (made non-conductive) . This is shown at cycles 22, 23 with buffer switch 16 open at t9 and tl3, for, e.g. 1 to 2 μs and 4 to 6 μs, respectively. Then, the signal accumulates in buffer capacitor 15, but is not relayed to the integrator 12. During these times, the circuit is free for administrative tasks, particularly for reading out and evaluating the signal from the previous cycle. During t9, reset switch 13 may be closed for til, spaced from earlier and later buffer switch closings t3, tl2 by t8 and tlO, for signal reset

During measurement in either measurement cycle 22 or 23 the switching scheme may be such that the reset switch 13 is closed (conductive) long enough for allowing a proper capacitor discharge (i. e. integrator reset, til, tl5, of e.g. 1 to 2 μs), and the buffer switch 16 is never closed (conductive) while the reset switch 13 is closed (conductive, til, tl5), for avoiding loss of measurement signal. This may be ensured by providing time spacings t6, tlO, tl6, of, e.g. 50 to 100 ns.

Within tl2 and tl7, buffer switch 16 is closed while reset switch 13 is open. The signal both accumulated meanwhile in the buffer capacitor and coming directly from the sensor element thus runs directly into the integrator 12 and is integrated there. Period tl2 in the variable length cycle may be of adjustable variable length as required, set by an external control signal, whereas tl7 may have a fixed duration of e.g., between 3 and 7 μs . At the end of tl2 and tl7, buffer switch 16 opens again and a measurement cycle is finished because no further signal reaches the integrator 12 which thus has a constant output ready for read-out and evaluation and later reset.

Within evaluation period tl4 (of, e.g. between 2 and 4 μs) between end of charge transfer (buffer switch 16 opens) and reset (reset switch 13 closes), the integrator output may be evaluated (by holding and sampling, storing a digital value, or the like) . This may be triggered by pulses of appropriate timing and duration. Plural evaluations of the integrator output may be made for averaging out noise. 27 shows short trigger pulses tl9 and t21 (e.g. 50 to 100 ns), spaced by t20 (of, e.g. 2 to 3 μs), all within tl4 and spaced from its beginning and end by tl8, t22, for triggering two evaluations of the same constant integrator output value.

It is noted that the evaluation period tl4 for the value reached in the preceding measurement cycle is done in parallel to accumulating charge again in the buffer capacitor 15 for the current measurement cycle. The first measurement cycle (22 in fig. 2) does not need an evaluation period because there is nothing to evaluate.

Generally speaking insofar, the invention can be described as a method for reading out, by a readout circuit, a sensor element that produces an electric signal in accordance with a phenomenon to be sensed, comprising the steps of alternatingly operating the readout circuit in an acquisition period and in an organization period. In the organization period (e. g. above tl3) the sensor element and a signal buffer are disconnected from the readout circuit which' may use this period for organization tasks (such as reading the integration value, resetting the integrator

(discharging capacitor 18), and others) and are reconnected to the readout circuit in a subsequent acquisition period (e.g. above tl7) for acquiring (integrating) in the readout circuit the meanwhile buffered and the newly incoming signal components.

The value read out from the integrator may be the voltage at the end of a measurement cycle, i. e. when the buffer switch is opened. Before being taken as an intensity measure for the received input quantity, it may, however, be compensated for noise contributions, e. g. by subtracting an earlier determined mean value of noise, generated by the electronic circuitry. The read-out results from plural integration cycles (before or after noise compensation) may be added up before being further evaluated.

In the following,- some exemplary value or value ranges are given: The buffer capacitor may have a value of below 200 or 100 or 50 pF. A bias voltage of -2.000 volt +/- 50% thereof may be applied to the cathode of a photomultiplier . Along the channel of the photomultiplier, the voltage may drop towards ground, but with a certain voltage gap (e.g. between 10V and 100V) kept both towards anode and cathode so that the electron avalanches finally reach the anode. The integrating capacitor may have less than 200, 100, 50 or 20 pF and its capacity may be different from, and particularly lower than, that of the buffer capacitor. A typical fixed measurement cycle 23a, 23b may be of 5 to 15μs overall duration, preferably 8 to 12 μs . A variable measurement cycle 22 may last minimum 2 to 5 μs and maximum the minimum value plus at least the duration of the fixed measurement cycle.

The sensor element may be for sensing electromagnetic radiation. It may be a more or less complex sensor element with built-in amplification, such as a photo multiplier, generally of dynode type, particularly a photo multiplier tube, a channel photo multiplier, a micro channel plate photomultiplier, a semiconductor photo multiplier an avalanche photo diode or a multi pixel avalanche photomultiplier respectively with driving circuitry determining the multiplying factor. The amplification may be caused by an avalanche of secondary electrons caused by earlier secondary electrons or by primary electrons generated by the incident radiation. Acceleration of the electrons may be caused by an electric field along their path. But likewise, non-amplifying radiation sensor elements such as photo diodes (PIN photo diodes) may be used.

Fig. 3a shows an electrical output characteristic of the integrator of the readout circuit. The abscissa may represent the quantity to be sensed, i. e. radiation intensity I, or a quantity representing the sensor element output, whereas the ordinate represents an output per measurement cycle (i. e. at the end thereof) , which may conveniently be a voltage U. In a low intensity range 31, 31', the input is too weak to cause an output above noise level N. The noise may come from the sensor element 1 itself or from the readout circuit. 32 and 32' is a monotonic and more or less linear range where the output follows the input. 33 and 33' represent saturation where the sensor element 1 itself and/or the readout circuit, particularly the integrator thereof, run into saturation.

For very weak intensities (namely max. one photon per cycle), a characteristics of output caused by a photon per measurement cycle (ordinate) over sensor element gain v (abscissa) could be drawn. It would look qualitatively the same as Fig. 3a so that it is not drawn a second time. When v starts at zero and increases, the electrical effect caused by said single photon (e.g. the avalanche caused by a photon) would likewise increase. At the beginning (i. e. near zero range 31) it may, however, be too weak to reach out of noise level N (sensor element noise and circuit noise). When further increasing, it will be distinctive from noise and output may be close to. linear (range 32, 32') . Further increasing v may lead to saturation, e.g. the integrator reaching its limits in view of driving conditions (range 33, 33' ) .

Reading out may have three different modes, namely a first ("photon counting") mode, a second

("threshold") mode, and a third ("integration") mode. For avoiding misunderstandings it is pointed out right here that all modes may use thresholds and integration, although their naming may suggest that only one of them respectively does so.

Fig. 3b shows partial characteristics 34 - 38 of these modes. The abscissa is logarithmic and shows the intensity of the signal to be sensed whereas the ordinate may also be logarithmic and shows an output value. For curve 34, it is "cps" (counts per second, ranging in the shown example from 0 to about 10^A5), whereas for 35 to 38 it may be the integration result of one or more measurement cycles (without a unit given in Fig. 3b) . It may be a voltage U or an electrical charge Q or an electrical current I, which, in turn, would be in the respective range proportional to the received intensity. It may be converted into one of the other mentioned electrical quantities (Q, I, U) or into a count value or photon numbers or intensity or a digital value.

The modes may selectively be used by setting respective sensor element driving conditions and signal evaluation conditions by a controller. The modes basically may correspond to different intensities of incoming radiation. Integration mode has been described above. The overall circuit or at least the analogue part thereof may for all modes be the same and may be driven qualitatively in the same manner, whereas the evaluation and interpretation of the circuit output (i. e. integrator output) may differ amongst the modes.

Photon counting mode is for the lowest intensity in which single photons are received that can individually be detected in a measurement cycle. It corresponds to curve 34 in Fig. 3b. This method can be used if the intensity is so low that per measurement cycle with the above mentioned circuit max. one (i. e. one or zero) photon is received. Each measurement cycle showing an output result above a certain threshold is taken as the representation of one photon. The number of measurement cycles showing an output result above said threshold is counted and represents the number of received photons. The settings (and particularly amplification/gain v of the sensor element 1) may be such that one photon brings the output in a measurement cycle above a first threshold Tl determined in view of noise and the considerations in relation to gain v explained with reference to Fig. 3a above.

Generally, in this mode a sensor signal is integrated over a certain period of time (measurement cycle 22, 23) and a count value is modified, preferably incremented by a fixed number, preferably 1, if the integration result in said measurement cycle exceeds said threshold Tl. This may be repeated for plural measurement cycles and may be continued over a time window of a fixed length, e. g., one second, so that a count rate is obtained.

As said, in this kind of detection, the detection threshold Tl may be set in accordance with the noise signal of the circuit and the sensor element. For determining the threshold, the mean value of noise and the standard deviation thereof may be determined from a sufficient number of independent noise signal sample measurement cycles. The threshold may be chosen to be n times the standard deviation of the noise above the mean value of noise, n being between 1 to 10, preferably between 4 and 6, preferably 5. Instead of a multiple of the standard deviation, a value depending on the maximum or average or sufficient deviation within a set of sample values may be taken, e. g. said maximum or average or sufficient deviation value added with, or multiplied by, a certain value.

The amplification/gain may be chosen such that in view of the threshold no photon or only a relatively small number of photons (e.g. 20% or less, 10% or less) gets lost from detection. During calibration, a low gain value may be set initially, which may be increased until measurement of a known intensity (with max. 1 photon per cycle) shows that all photons caused an output above the threshold (i. e. an increase in amplification does not lead to an increase in counts because all events are already above the threshold) .

If a measurement cycle is 10 μs, the maximum count per second in this mode is 100.000 (= ls/lOμs). But for being with sufficient accuracy at the above assumption (i. e. max. 1 photon per cycle) in view of the statistical nature of photon incidence, this mode may be applied only for count rates lower than 100.000 per second for excluding with sufficient confidence that two or more photons per cycle are received. A factor 3 or 5 or 10 below the theoretically maximum count rate may be a realistic upper limit for photon counting mode, which would be 33.333 or 20.000 or 10.000 counts per second at 10 μs measurement cycle which may be taken as a second threshold T2 as an upper limit for using photon counting mode. Counting mode may then cover an input intensity range of factor 10^Λ4 or more. But likewise, a value to be compared with the output from one measurement cycle may be taken as a second threshold T2. If a photon produces a certain output, a value above said certain value may be taken as second threshold T2. If T2 is exceeded by the actual output of a measurement cycle, this may be a measure for that two or more photons have been received, so that photon counting mode is no longer appropriate .

It needs to be remarked that counting mode is only applicable to sensor elements with an internal gain v high enough to bring for a single received photon the output above noise level. These Sensors can be used in all three modes, whereas sensors with small or no internal amplification can be used in integration mode or might be used to some extend in threshold mode. The internal gain v is preferably adjustable.

Threshold mode can also be used for the above- mentioned lowest intensities such as the detection of single photons per cycle, but is also suited for higher intensities and may follow, on an intensity scale, the photon counting mode. It corresponds to curve 35 in Fig. 3b. Plural photons may be received in one measurement cycle, and the sensor element output signal caused by each of these photons within one cycle is integrated by the usual operation of the circuit. Gain of the sensor element and the threshold of the electronics may be equal to the photon counting mode .

In contrast to the photon counting mode where the output of each read-out is 0 or 1 photon, the threshold mode evaluates the value (current, charge or voltage) read out in each measurement cycle quantitatively. This quantity is noise-affected. For obtaining a more precise result, the noise contribution to the output signal, e.g. the mean value of noise, may in threshold mode initially be determined and during measurement be subtracted from the integrator output before it is further used. Only read-outs above a given threshold may be used. If the read-out in one measurement cycle is below the given threshold the read-out may be taken as zero. If the read-out exceeds the given threshold, the noise contribution of the electronics (sensor, circuit) will fully or partially (30% to 70%) be subtracted from the measured read-out, and the result will be further processed. Through this, noise contribution in a range of still relatively low intensities is eliminated so that the obtained result is more precise.

The output at the end of the measurement cycle, reduced for noise contribution as said above, may be caused by, and thus may represent, plural photons, and the obtained value may be converted into number of received photons by using a conversion constant representing an average electrical effect per photon under the given driving conditions. This may be done on the digital side by computation or on the analogue side. The result from this mode may fit seamless to the result from photon counting mode (characteristics 34) .

Generally, in threshold mode a sensor signal is integrated over a certain period of time, such as an above mentioned measurement cycle, and the integration result thereof can be converted (with reference to noise data as said above) into some kind of intensity or count number. For low intensities, particularly the overlap region with counting mode 34, the outputs of the individual measurement cycles must be added over a time window (e. g. 1 s) covering many measurement cycles for obtaining a meaningful result. Accordingly, with the value obtained from the recent measurement cycle an already existing intensity or count value is modified, preferably by adding the recent value to the existing value. This may be continued over a time window of, e. g., one second, so that an integrated intensity or a count rate is obtained. This way of adding results from many measurement cycles can be made through the entire intensity range suitable for threshold mode. The more cycle results are added, the lower is the influence of statistical variations of photon incidence.

But likewise, for higher intensities the output of a single or some few measurement cycles may be taken and may appropriately be processed, particularly added. Within integration mode the number of cycles to be added may be altered according to a defined policy from higher for low intensities to lower down to 1 for high(er) intensities. The conversion into another or final quantity can also be made after intermediate or raw values from several cycles were added, e.g. in the above mentioned time window.

The electronics (analog circuit, integrator) itself may have a linear range between noise and saturation of about 10^A3 or 10^Λ4. It represents the upper 3 to 4 orders of magnitude in curve 35 in Fig.3b, whereas in the lower intensity regions (particularly overlap with counting mode 34) the output may in many cycles be zero (no photon detected) or short above noise level (one or few photons detected) .

Threshold mode may therefore cover an input intensity range of a factor of about 10^Λ8 or more, perhaps 10^Λ4 thereof being overlapping with counting mode on the low intensity side of the range.

Integration mode is used when photon intensity is again higher. It corresponds to curves 36 to 38 in Fig. 3b. It may be distinguished from photon counting mode and from threshold mode by a modified (lower) gain setting. A more or less steady stream of photons is received. Integration is made as described above. The output may be given as some kind of quantitative measure Q. It may also be converted into cps, if deemed useful. In integration mode, plural different settings for sensor element driving and signal evaluation may be adjustable, e. g. in accordance with sensed signal intensity or operator or other input, said settings leading to the different curves 36 to 38, distinguished by different gain settings or high voltage settings of the sensor or integration periods (i. e. length of fixed length measurement cycle 23) or integration capacities. For each of the curves 36 to 38 a particular conversion constant may be used for converting the obtained signal into a desired quantity.

The three characteristics 36, 37, 38 may correspond to different sets of driving and evaluation conditions. For example, the sensor gain v may for curve 36 be higher than for curve 37, for which it is in turn higher than for curve 38. Likewise, the measurement cycle may have a decreasing length or integration capacity may increase from 36 to 38 for avoiding saturation.

Further, depending on circumstances, each of the modes (counting, threshold, integration) may be used for it alone in a sensor device, or one of the modes may not be used while nevertheless covering the input intensity range continuously with the other modes. Integration mode may immediately follow counting mode without threshold mode inbetween, or counting mode may be left away and the low intensity range is covered by threshold mode. Likewise, threshold mode may be continued up to high input signal ranges, thus avoiding integration mode, possibly with differing settings as shown with curves 36 to 38 for differing ranges within integration mode.

In all mentioned modes, the operation of the analog circuit part and/or the driving and switching scheme and read-out of the circuit may be the same, e. g. as shown in Fig. 2. But the evaluation of the readout value, particularly on the digital side in circuit 19a, is then different amongst the modes.

Thresholds T2, T3, T4 and T5 may be used as criteria for mode switching as shown. It is pointed out that these thresholds are shown on the abscissa of Fig 3b. But they may be implemented by monitoring the count rate or the integration output in a measurement cycle and may thus substantially be compared with a counter output or with the ordinate value of the characteristics of Fig. 3a (to be read on output over incoming intensity). Insofar, Fig. 3b shows the intensity equivalent of the used thresholds on the abscissa, but not necessarily an actually used threshold value. T3, T4 and T5 may have the same quantitative value (in the analog circuit upper threshold towards saturation at the integrator or lower threshold towards noise in the analog circuit) , and their different meaning comes from the different sets of driving and evaluation conditions in which they are used.

The integration mode(s) may together cover together an input intensity range of a factor of about 10^Λ6 or more. Together with the other modes, thus, a range of 10^A12 (because of overlap) and more is covered.

Generally speaking, the above mentioned conversion constant, noise level, thresholds and sensor element amplification of the sensor element itself and of the readout circuit may be determined under appropriate or known conditions, also repeatedly or periodically or triggered by an operator. For determining overall sensor noise, the sensor element input may be shaded and the resulting signal may be evaluated. For determining circuit noise, the sensor element may be disconnected or gated from the circuit supplied with an inverse bias, and then the circuit output is evaluated. For determining circuit amplification or sensor element amplification, one or more reference intensities may be applied to the sensor element, preferably by a built-in reference source, and the resulting signal is evaluated.

Switchover amongst the modes may be made during measurement, particularly between two cycles thereof. It may be made automatically. Mode setting/switching may be made with reference to certain thresholds in accordance with one or more of the intensity of the sensed radiation, the integration output, a count value or a count rate so that each radiation intensity range receives an appropriate readout mode. A hysteresis characteristics may be used for mode switching. But mode switching may also be made in accordance with external input, e. g. from some kind of higher level control or from an operator.

The readout method may also require controlling the operation conditions of the readout circuit (e. g. setting the duration of a fixed length measurement cycle 23 to an appropriate value, deciding integrator capacitance 18) and/or the driving conditions of the sensor element, particularly a possible amplification/ gain thereof. The driving conditions may be controlled in parallel to or in accordance with the desired readout mode mentioned above. Mode switching may insofar involve change of the driving conditions of the sensor element. Particularly, the gain of the sensor element may be switched, possibly in accordance with mode switching. The higher the intensity is, the lower may the gain be set.

Gain switching of a channel photo multiplier or of a photo multiplier tube or avalanche photo diode may be made by changing the cathode voltage against the anode, or changing the voltage between a downstream portion, particularly the downstream end of the channel (respectively dynode stages) and the anode, or both. This is an aspect of the invention (to be seen as gain switching method and as a sensor element) that may also be used independent of the other features described so far and/or below. The sensor element comprises an electron-emissive cathode, a channel along which a voltage drops for generating secondary electrons, and an anode for collecting electrons. Voltage applying means are provided and adapted to changing the voltage between a downstream portion, particularly the downstream end, of the channel and the anode by an amount different from the amount by which the voltage between the cathode and the anode is changed. For example, a variable and adjustable voltage source may be connected between the anode and said downstream portion, while the voltage between anode and cathode is kept constant or is independently (differently) adjustable. Through this, the voltage drop along the channel can be adjusted by changing the potential of the downstream portion/end in relation to the cathode without requiring a change of the cathode potential itself.

Mode switching of the overall circuit may also involve changing the circuit parameters and/or changing circuit driving parameters, particularly changing the feedback capacity of the integrator by selectively connecting/disconnecting parallel feedback capacities. Likewise, the integration period (measurement cycle) may be changed, e. g. shortened or made longer, for keeping the output signal in a reasonable and linear range.

Mode switching may also involve change of the evaluation algorithm, which may be executed on the digital side.

Generally speaking, the overlap of modes on the intensity range (abscissa in Fig. 3b) can be used for calibrating the ranges against each other by comparing within an overlap region the results obtained from the overlapping modes and making adjustments for bringing the results into agreement. Evaluation of a certain measurement result during a measurement cycle can be made according to any of the applicable modes, particularly in digital component 19a. So, the result from the different evaluations can be compared, and settings on the evaluation side (proportional factors) and/or on the driving side (sensor gain, integration parameters) can be changed for equalizing them.

Very low radiation intensities can be set for calibration purposes with good accuracy. With this, counting mode (characteristics 34) can accurately be calibrated. From there, using an intensity somewhere in the overlap between counting and threshold mode, threshold mode (characteristics 35) can be calibrated by comparing its results to calibrated counting mode results and adjusting threshold mode settings such that the results from the two modes coincide. From there, using an intensity somewhere in the overlap between threshold mode and integration mode, integration mode (characteristics 36) can be calibrated by comparing its results to calibrated threshold mode results and adjusting integration mode settings such that the results from the two modes coincide. From there, using an intensity somewhere in the overlap between two integration modes with different settings, the next integration mode (characteristics 37) can be calibrated by comparing its results to the already calibrated integration mode results and adjusting the other integration mode settings such that the results from the two modes coincide .

The above describes very low intensities as starting point for the calibration cascade. But any starting point is suitable where a well-known radiation intensity can be adjusted, and the cascade can go both in direction towards higher and towards lower intensities .

For example, the low intensity end of curve 35 and curve 34 below a count rate of le+4 in Fig. 3b can be used to calibrate the read-out value of the threshold method with the count rate obtained in the photon counting mode. Photon counting mode and threshold mode can be evaluated for each read-out simultaneously in parallel. For low intensities below 1000 cps the photon counting mode may be more accurate since the absolute number of photons is given.

Generally speaking insofar, the invention teaches also a method of calibrating a readout circuit capable of reading out a sensor element in plural readout modes (such as photon counting mode, threshold mode, integration mode as mentioned above) , the method comprising a first calibration step of calibrating the circuit in a first mode (e. g. counting), and a second calibration step of calibrating a second mode of the circuit (e.g. threshold, integration) with reference to results in the calibrated first mode (e.g. counting, threshold) .

The very first calibration step can be absolute, i. e. making calibration with reference to a known input quantity and adjusting settings such that the output in said very first mode matches the known input quantity. The subsequent mode calibrations may be relative by making settings such that the output of the mode to be calibrated matches the output of the already calibrated mode.

The individual steps of each of the readout modes downstream of the integrator may be made on the digital side after A/D conversion of the integrator output value at the end of a measurement cycle. Two or more of the above-mentioned readout modes may be made in parallel on the digital side for obtaining parallel results from the parallel-running modes, and according to predetermined criteria one of the obtained results may be selected as output. Mode switching involves then a selection from plural possible outputs.

The term "mode" is used insofar not only for distinguishing qualitatively different approaches (such as above counting, threshold, integration mode, characteristics 34, 35, 36), but also for distinguishing quantitatively different approaches

(such as the above different integration mode 36, 37, 38) .

The circuit may have self-calibrating means, amongst them a radiation source and a control means for adjusting appropriate settings during calibration, making appropriate evaluations, and determining settings for ordinary measurement operation. These settings may be one or more of gains of the sensor, thresholds for mode switching, thresholds for photon counting mode or threshold mode, criteria for change of feedback capacity. Calibration may be made from time to time automatically or triggered externally. The newly acquired settings may replace older ones .

Various of the above-described method aspects can also be used with a conventional readout circuit. For them, particularly a readout circuit without a buffer capacitor 15 and/or without a buffer switch 16 may be used. Such a conventional circuit may otherwise be built as shown in figs. 1 and 4b and as described above. The method aspects that can be used together with such a conventional circuit are the above described readout methods (photon counting mode, threshold mode, integration mode) , the way of switching amongst said modes, and the way of calibrating the modes against each other. The conventional circuits show reduced detection accuracy as explained above. But as far as this is sufficient, the mentioned method aspects still exhibit their advantages of allowing a detection with a single circuit over a wide range and well adapted to respective input quantities, and allowing a precise calibration.

Claims

1. Sensor readout circuit (2) comprising an input terminal (11) for a sensor element (1) producing an electric sensor signal, an integrator (12) for integrating the sensor signal, a reset means (13) for resetting the integrator, and a control circuit (14) for repeatedly operating the reset means after an integrating period,

characterized in comprising

a buffer capacitor (15) connected to the input terminal and comprising a buffer switch (16) operable by the control circuit for connecting the input terminal with the integrator input within the integration period and disconnecting it therefrom during integrator reset.

2. Sensor readout circuit (2), particularly according to claim 1, comprising an input terminal (11) for a sensor element (1) producing an electric sensor signal, an integrator (12) for integrating the sensor signal, a reset means (13) for resetting the integrator, and a control circuit (14) for repeatedly operating the reset means after 'an integrating period,

characterized in that

the control circuit is adapted to operating the reset means for establishing integrating periods of 30 μs or less.

3. The circuit of claim 1, wherein the integrator comprises an operation amplifier (17) and a feedback capacitor (18) preferably of a capacity smaller than that of the buffer capacitor, and the input terminal is connected to the feedback loop of the integrator.

4. The circuit of claim 3, wherein the reset means comprises a reset switch connected in parallel to the feedback capacitor.

5. The circuit according to one or more of the preceding claims, wherein the integrating period is 20μs or less and is preferably higher than 100 ns .

6. The circuit according to one or more of the preceding claims, comprising an A/D-converter (10) for converting the integrator output signal or a signal derived therefrom into a digital value, wherein the control circuit comprises an evaluation circuit (19) for receiving and evaluating the A/D converter output signal.

7. The circuit according to one or more of the preceding claims, adapted for the connection to a sensor fox electromagnetic radiation, preferably with internal amplification, preferably a photomultiplier tube (1) , channel photomultiplier, micro channel plate photomultiplier, avalanche photo diode, multi pixel avalanche photo diode, semi-conductive photomultiplier and/or photodiode, PIN diode.

8. The circuit according to one or more of the preceding claims wherein the evaluation circuit (19) comprises a sampling circuit for sampling the integrator output at the end of the integrating period before the reset.

9. A sensor comprising a sensor element (1) and a readout circuit (2) according to one or more of the preceding claims .

10. The sensor of claim 9, comprising a sensor driving circuit (3), preferably controlled by the control circuit of the readout circuit.

11. The sensor of claim 9 or 10, wherein the sensor element is a photomultiplier tube, channel photomultiplier, micro channel plate photomultiplier, avalanche photo diode, multi pixel avalanche photo diode, semi-conductive photomultiplier and/or photodiode, PIN diode.

12. Method for reading out, by a readout circuit, a sensor element that produces an electric signal in accordance with a phenomenon to be sensed, comprising the steps of alternatingly operating the readout circuit in an acquisition period and in an organization period,

characterized in that

in the organization period the sensor element and a signal buffer are disconnected from the readout circuit and are re-connected to the readout circuit in a subsequent acquisition period.

13. The method of claim 12, comprising the step of integrating the signals from the sensor element and from the signal buffer in the acquisition period and evaluating the integration result and resetting the integrator in the organization period.

14. The method of claim 12 or 13, wherein the^' readout period comprising the acquisition period and the organization period is less than 30 or 20 μs .

15. Method of operating a sensor, comprising the steps of obtaining plural sensor measurement cycle values in a definable number of measurement cycles, and adding the plural cycle values for obtaining a final value.

16. The method of claim 15, wherein the cycle values are obtained by the method according to one or more of the claims 12 to 14.

17. Method of operating a sensor capable of being operated in two, three or more different modes, the modes respectively being defined by a set of sensor driving settings and/or of sensor readout settings, the method comprising the steps of evaluating the sensor output signal, and switching over amongst the sensor driving settings and/or the sensor readout settings to be used for the various modes in accordance with a criterion, said criterion preferably being the result of an evaluation of the sensor output signal.

18. Method of claim 17, wherein one of the operation modes is the readout method in accordance with one or more of the claims 12 to 14.

19. Method of claim 17 or 18, wherein one of the operation modes is integrating a sensor signal over a certain period of time and modifying a count value if the integration result in said period of time exceeds a threshold.

20. Method of claim 17, 18 or 19, wherein one of the operation modes is integrating the sensor signal over a measurement cycle and evaluating the output only if the read-out exceeds a certain threshold.

21. Method according to claim 20 wherein if the read-out exceeds the given threshold the mean value of noise level will be subtracted from the measured readout.

22. Method of claim 21 wherein the threshold is determined by measuring the mean value of noise and a deviation value, preferably the standard deviation of the noise of the electronics, and determining the threshold to be a value corresponding to the mean value of noise added with n times the standard deviation of the noise of the electronics, obtained from a sufficiently high number of measurement cycles, wherein n may be a value between 1 and 10.

23. Method of one of the claims 17 to 22 wherein the gain value of the sensor, preferably a photomultiplier is set so that in view of the threshold no or less than 20% of the photons gets lost from detection.

24. Method of claim 17 or 18 wherein each readout value is subtracted by the mean value of noise level of the electronics for the given read-out time.

25. Method of one of the claims 17 to 24, wherein one of the operation modes is integrating a sensor signal over a certain period of time, converting the integration result into a count number and modifying a count value in accordance with the obtained count number.

26. Method of calibrating a readout circuit capable of reading out a sensor element in plural different readout modes, comprising a first calibration step of calibrating the readout of the readout circuit in a first readout mode of the readout circuit, and a second calibration step of calibrating the readout of the readout circuit in a second readout mode of the readout circuit with reference to results in the calibrated first readout mode.

27. Method according to one or more of the claims 17 to 26, wherein a readout circuit as described in one of the claims 1 to 8 is used, but said readout circuit not having said buffer capacitor (15) and preferably not having a buffer switch (16) .

28. A method of gain switching of a sensor element comprising an electron-emissive cathode, a channel along which a voltage drops for generating secondary electrons, and an anode for collecting electrons, characterized in comprising the step of changing the voltage between a downstream portion of the channel and the anode by an amount different from the amount by which the voltage between the cathode and the anode is changed.

29. A sensor element comprising an electron- emissive cathode, a channel along which a voltage drops for generating secondary electrons, and an anode for collecting electrons, characterized in comprising voltage applying means adapted to changing the voltage between a downstream portion of the channel and the anode by an amount different from the amount by which the voltage between the cathode and the anode is changed.