WO2004077340A1

WO2004077340A1 - Substrate multiplexing with active switches

Info

Publication number: WO2004077340A1
Application number: PCT/NO2004/000057
Authority: WO
Inventors: Jon Nysaether
Original assignee: Idex Asa
Priority date: 2003-02-28
Filing date: 2004-02-27
Publication date: 2004-09-10

Abstract

This invention relates to a sensor module for measuring structures in a surface, especially a finger surface, comprising number of sensor elements being located in chosen positions in a common surface, the sensor elements being arranged in an essentially linear array and an external reference potential, the sensor elements also being coupled to an electronic circuit, said electronic circuit being adapted to measure the magnitude of the capacitance or impedance between chosen sensor elements and said external reference potential at chosen points of time, said electronic circuitry comprising input means including at least two amplifier circuits, each amplifier circuit being coupled to a set of sensor elements including at least two elements for amplifying signals therefrom and transmitting it to said electronic circuitry. Thus a multiplexing of the measured signal is achieved by simultaneous measuring of different parts of the surface in parallel sequences.

Description

SUBSTRATE MULTIPLEXING WITH ACTIVE SWITCHES

This invention relates to a sensor module being capable of sensing the topological pattern of a surface in direct contact with it, e.g a fingerprint.

The market of biometrics is evolving rapidly. However, for biometrics to penetrate the consumer market, there are strict requirements with respect to e.g. sensor price, compactness, fingerprint image quality and power consumption. Capacitive fingerprint sensors are one of the most promising technologies for realising compact and low cost fingerprint sensors for the consumer markets, and several concepts have been proposed over the past few years. The sensor concepts roughly divide into two categories: matrix sensors, where the fingerprint is placed onto a two-dimensional sensor surface, and scanners or sweep sensors, where the user is required to pull his or her finger across the sensor to capture an image.

US Patent 6,069,070 describes a typical matrix type, AC capacitive fingerprint sensor. The sensor is basically a silicon chip (IC) which is furnished with a 2-D matrix of sensor elements ("pixels") as well as amplifiers and other circuitry. A drive electrode located on the sensor package, outside the active sensing surface, is used to couple an AC voltage into the finger. The AC signal then penetrates the finger and couples through a dielectric layer to sensor elements (pads) on the surface of the sensor. The sensor pads are coupled to amplifying circuitry in the silicon chip.

Another example of a matrix sensor is shown in US 5,325,442. The sensor elements are here organized in rows and columns, and a specific element can be addressed by selecting the relevant row and column. Each row is thus coupled to a drive circuit which can be used to apply a voltage to the sensor element, and each column is coupled to a sense circuit. For each sensor element there is also a transistor switch which can be used for switching the connection to the drive circuit. By addressing one row at a time and sample each sense element each time a new row is addressed, the full matrix of sensor elements can be accessed to obtain a full 2-D image of the finger surface.

These sensor configurations however has several disadvantages: For obvious reasons, the active surface of a matrix sensor needs to be as large as the portion of the finger that shall be imaged, in other words in the order of 100 mm². Because the price of silicon ICs increases proportionally with the chip area, such a large chip may be forbiddingly expensive for many consumer applications.

Finge rint scanners, where the user is required to pull his or her finger across the sensor, do not have to have the same length as the fingerprint, and may therefore be smaller and more cost-effective. Because of their reduced size and lower price, scanning sensors may be a better choice for typical mass consumer applications. Scanners require only a very limited number of sensor lines, and while the width of the sensor still has to match the width of the fingerprint, the sensor length may be as low as a few mm or even less. The "linear" arrangement of sensor elements also gives more flexibility in designing the interface between the sensor and the finger - it does not necessarily have to be a flat surface like matrix sensors often are.

Patent application WO01/99035 shows an example of a scanning sensor where the integrated circuit with amplifiers etc (ASIC) is mounted on the back side of a substrate. The top side of the substrate serves as the interface between the finger and the sensor. The substrate is for instance silicon, ceramics or glass. The top side of the substrate is equipped with a number of conductive pads for sensing the capacitance, and conductive vias are made through the substrate to connect each sensor pad or element with a corresponding input pad on the ASIC. The sensor pads are covered with a dielectric material. The described sensor comprises an essentially linear array of sensor elements with one full, single line and additional sensor elements for determining the speed of the finger. Alternatively two lines with "staggered" sensor elements can be used, and the readings from the two lines can be used for obtaining the speed. The speed measurement is needed for building up a 2-D image based on the readings from the sensor. Examples of other scanning sensors are given in WO 97/58342, WO01/99036, US 6,289,114, and EP 0 735 502 A2.

This substrate based sensor in WO01/99035 has several advantages: The possibility to "fan in" the signal tracks on the substrate makes it possible to decouple the size of the ASIC from both the width and length of the fingerprint to be captured. This makes it possible to design a much smaller IC chip and thus to save cost. In addition, the substrate, which is far less expensive to make per area, serves as a mechanical and environmental protection for the vulnerable ASIC on its back side. The sensor has a very low profile, and if supplied with e.g. BGA balls it does not have to be further packaged before mounting it onto a motherboard, e.g in a phone.

The sensor also has the advantage that drive electrodes for stimulating the finger with an AC signal can be integrated directly on the surface of the substrate. These drive electrodes may be coupled to ground through an ESD protecting semiconductor device, so that any ESD discharge from the finger will go to the electrode rather than to the sensor pads.

The device in WO01/99035 however requires one ASIC input channel, and one amplifier, for each sensor pad on the top surface. An amplifier with adequate analogue filtering normally consumes a significant space on the ASIC. In addition, the i/o pad and related ESD protection circuitry consumes space. With so many input channels, it will thus be difficult to reduce the size of the chip below a certain limit. This may make it difficult to reach the extremely low production costs needed for a mass produced sensor. The concept in WO01/99035 also requires that a large number of routing tracks, in principle one for each sensor element in the line, be connected to the ASIC, either by means of via holes through the substrate or by wire bonds or TAB (tape automated bonding) on the sides. This has several disadvantages: For through-substrate vias, the employed substrate technology must allow for a large number of vias within a relatively small area. This makes the substrate process complicated or may even exclude the use certain substrate materials or processes.

For wire bonding or TAB, the large number of connections is unpractical and adds extra cost to the sensor. In addition, this type of connections must be protected e.g by a glob-top and thus adds height around the rim of the sensor. This is unpractical from a finger movement point of view, especially if the long side of the sensor must be used to connect all channels.

It is therefore an object of this invention to provide a sensor module in which the number of individual amplifiers on the ASIC is significantly reduced. This will lead to a simpler substrate and reduce the size of the ASIC, thus reducing the production cost of the sensor.

Moreover, it is an aim of the invention to perform this reduction of channels using a form of sensor element multiplexing where the sensor elements of the essentially linear sensor are grouped in a number of sets, the elements in each set being coupled to the same amplifier and a control signal being used to access or activate the sensor elements in a set in a given sequence, one at a time. The time signals from the various amplifiers are read out in parallel. Within a certain time frame, namely the time it takes to access all sensor element in a set, and by associating the known positions of the sensor elements to the activation times, an essentially linear image of the fingerprint can be obtained.

The present invention aims primarily at a scanning sensor principle where a finger is swiped across the sensor in a direction essentially perpendicular to the line of sensor elements, and where linear images of the finger surface are acquired at chosen intervals. The invention is characterized as described in the independent claim.

According to a preferred embodiment of this invention the sensor is a scanner with an essentially linear array of sensor elements fabricated on a substrate. The elements are organized in groups, where the elements in a group share a common control or activation electrode that can be used to "activate" or excite the elements in this group only. One element from each group is coupled to a common signal track, connected to an input channel of an ASIC or in a discrete electronic system. This is done in such a way that a signal current or voltage on said signal track is received only from the element belonging to the group with an activated electrode, and no signal is received from the other elements belonging to other (unactivated) groups.

The sensor elements that are coupled to the same amplifier are termed a "set" of sensor elements.

It should be clear from the above that each sensor element in the essentially linear array belongs both to a "group" and to a "set" of sensor elements, and that no sensor elements belong to both the same group and the same set. By specifying the group and set, a single sensor element is thus identified, much in the same way that the column number and row number identify an element in a 2-D matrix. In the following discussion, the words "group" and "set" will by analogy respectively be termed "row" and "column". The use of the expressions "rows" and "columns" does however not imply that the sensor elements are physically organized in a matrix fashion on the substrate or on the ASIC, but rather that each element in the essentially linear array has a double connection, with one connection to an activation or control electrode and another to a sense circuit.

According to a preferred embodiment the geometry of the activation electrode is so that there is a capacitive signal coupling from said electrode to the sensor elements of said row when an AC current or voltage is applied to said activation electrode. From each sensor element in the row there is also a capacitive signal coupling to a signal track leading to one input channel of the ASIC. Finally, from each sensor element there is a capacitive or galvanic coupling to an external ground or other reference potential through the fmger, so that the signal received at the ASIC input channel is modulated by the presence of a fingerprint ridge or valley situated directly above the sensor element.

The invention is described below with reference to the accompanying drawings, illustrating the invention by way of examples. Figure 1 illustrates schematically the circuit associated with one sensor element. Figure 2 illustrates a schematic circuit according to the invention. Figures 3 and 4 illustrates longitudinal sections of two different embodiments of the sensor. Figure 5 illustrates a sensor according to the invention as seen from above. Figure 6 and 7 illustrate cross sections of the top layers of the sensor.

Figure 8 illustrates a longitudinal section of an embodiment of the sensor. Figure 9 illustrates a possible arrangement of the sensor elements in "rows" and

"columns". Figure 10 illustrates schematically the circuit associated with one sensor element for an alternative embodiment.

Figure 11 and 12 illustrate schematic circuits of alternative embodiments of the invention. Figure 13 illustrates a cross section of the top layers of the sensor for an alternative embodiment. In one preferred embodiment, detailed in figure 3, 5, 6 and 7, the invention consists of a substrate 2 which maybe fabricated from e.g silicon, glass, ceramics, flexprints or a laminate board. On the top surface of the substrate, over which the finger to be imaged 1 shall be pulled, a number of sensor elements 5 are defined, e.g as conductive pads made in a thin film or thick film process. The sensor elements are organized in an essentially linear array configuration. The sensor elements may preferably be covered with a dielectric 8 (such as SiO2, SiN, polyimide, epoxy or a ceramic), or be in galvanic contact with the finger. The essentially linear array may for instance be arranged as a single line with additional sensor elements or groups of elements 6 for detecting the speed and/or the direction of movement, including rotation, of the fmger across the sensor. These movement measurements may either be used for correcting the image with respect to variances in finger speed, pull direction or rotation, or e.g. to move a cursor across a screen ("mouse").

Preferably the sensor is equipped with a number of activation electrodes 11 and signal tracks 10 leading to the input channels of the ASIC 4. Said electrodes and tracks may for instance be fabricated in a metal layer directly below the sensor elements, and separated from the latter by a dielectric material to achieve a capacitive coupling between the sensor elements and said electrodes and tracks.

The sensor elements 5 are preferably grouped in a number of rows, each with at least one activation electrode, which is shared between the elements in the row.

Each signal track 10 is capacitively coupled to several sensor elements 5, preferably one element from each row.

In a preferred embodiment the surface of the sensor is at least partly covered with a metal electrode 7 or other conducting plane held at a reference potential, e.g ground. This electrode may be in direct contact with an applied finger 1 or covered by a dielectric material.

In a preferred embodiment the substrate further contains a number of via holes 3 through which connections from the signal tracks 10 and the activating electrodes 11 are taken through the substrate 2 to the back side. The via holes 3 are generated by laser or mechanical drilling or an etching process and must subsequently be electrically isolated from the substrate, if the latter is conductive. This may be done by e.g. oxidation or deposition techniques. A conductive material may be deposited in the vias to make electrical contact through the substrate. Preferably the holes are sealed. If necessary, additional routing is included between the bottom end of the via holes and the substrate bonding pads of an ASIC 4 which is mounted on the back side of the substrate (e.g flip-chip or wire bonded). Preferably, the interconnections between the various parts of a signal track or a control electrode are connected in the conductive layers on the top surface of the substrate, so that the necessary number of through-substrate vias is minimized. These connections 15 may for instance be carried out by use of electrical vias between different conductive layers on the top surface. There may alternatively be one through substrate via 3 for each sensor element, and the connections 15 be carried out in routing layers on the back side of the substrate.

In an alternative embodiment, shown in figure 4, there are no through-substrate vias 3, but connections from the ASIC 4 to the signal tracks 10 and activating electrodes 11 on the substrate are made on the edge of the substrate 2, e.g. using wire bonds 9. Because of the limited number of necessary connections, wire bonding may in some cases be a more practical and less expensive solution than using vias 3. A drawback with wire bonds is that they extend above the top surface of the substrate, making it difficult to obtain a completely flat top sensor surface which is advantageous from a ergonomic point of view. As visualized in figure 8, this drawback may be solved by etching a step 26 on the substrate, e.g. on one of the edges, so that the sensor elements are situated on the highest level and the wire bond pads on the lower level. The etching of such a step may for instance be carried out using wet, anisotropic etching (KOH or TMAH) and subsequent passivation (e.g oxidation) on the substrate. The tracks to connect the two levels and the wire bonding pads on the lower level may be fabricated by ordinary thin-film processes using a photoresist that conformally covers the steps.

The substrate and ASIC may for instance be placed on a common lead frame 20 and connected by wire bonds 9, e.g. directly or via the lead frame. The assembled lead frame may then be molded in plastic 21.

In case of silicon or any other conducting or semiconducting substrate type, the substrate 2 is preferably at a fixed potential. The ASIC 4 (or an equivalent electronic signal conditioning system) contains a number of amplifiers 16 to amplify the signal associated with each signal channel, as well as other signal conditioning circuitry. Associated with each amplifier channel there may be a band pass filter for filtering the relevant AC signal frequency. Finally, the ASIC contains at least one activation circuit for feeding the activation electrodes 11 with an appropriate signal. The activation circuit may for instance have the form of an AC drive circuit coupled to n output channels through a multiplexer. The number n is here for instance equal to the number of activation electrodes 11 on the substrate. The multiplexer may for instance be programmed so that within a certain time frame corresponding to a limited movement of the finger (e.g less than 50 μm), all elements in a column are activated through their associated electrodes, one row at the time. For each time a new electrode is activated, each of the input channels of the ASIC are monitored at least once, so that a full reading of a row is performed. In this way all sensor elements (all columns, all rows) in the essentially linear array are accessed within said time frame. This gives a complete "line scan" of the finger, in addition to optional readings from the sensor elements used for detecting finger speed. Speed detection is carried out by comparing and correlating the time histories from pairs of sensor elements separated by a known distance in the direction of finger travel. Based on the detected speed, the line scans are subsequently used for reconstructing a correctly scaled 2-D image of the finger.

Preferably, a column of sensor elements consists of a number of neighbouring elements. Such an arrangement is illustrated in figure 9 for a sensor with 16 elements in each column CI, C2, C3. Preferably, the elements are so organized that two neighbouring elements in the same column are always accessed directly after one another. If the output from a single sensor amplifier channel is low-pass filtered, this will then be equivalent to applying a geometrical low-pass filter to the image in the direction of the linear array. This is advantageous from a noise reduction and image quality point of view. However, the geometrical organization of the rows and columns of sensor elements can be carried out in a number of different ways within the scope of this invention.

A sensor with a line of 256 elements at 50 μm pitch may for instance have 8 rows of 32 sensor elements. In addition there may for instance be two extra rows 6 of 32 elements used for speed measurements and pointer functionality as described in patent application WO01/95305. This sensor requires an ASIC with 32 input channels and amplifiers, and 10 output pads for connecting the different activation electrodes. In other words, only 42 connections between the ASIC and the top side of the substrate are required, instead of 320 channels for the principle described in WO01/99035. In addition comes ground connection and possibly other individual connections. Sensor elements belonging to different functionalities may share the same activation electrode.

A schematic version of a preferred measuring principle is shown in figure 1. An AC voltage signal from the activation electrode 11 couples to the sensor element 5. In the case shown in figure 1, the input signal current to the ASIC amplifier 16 is determined by the capacitor 12 in series with the total capacitance from the sensor pad to ground potential. Said total capacitance is given by 17 in parallel with the series capacitance of 13 (through the sensor dielectric 8) and 14 (through the finger ridge or through an air gap in a valley). The additional series impedance 18 through the finger to an external potential is for this discussion assumed to be negligible. Because all other elements are fixed, the magnitude of the input signal current will change dependent on the magnitude of the capacitance 14, which varies depending on whether there is a finger ridge or valley present. The signal current can be amplified, filtered and demodulated (e.g. synchronously) by the ASIC. In other words, when an AC voltage is applied to the activation electrode 11, there will be a capacitive flow of current from the activation electrode 11 to the signal track 10 through the sensor element 5. Direct coupling between 10 and 11 is assumed to be negligible.

When there is an air-filled fingerprint valley situated directly above the sensor element 5, the impedance from the sensor element to an external potential (at the electrode 7) through the finger 1 will be practically infinite.

On the other hand, when there is a fingerprint ridge present, there will be a much lower, finite impedance from the sensor element to the external potential through the finger. This will lead to a reduction of the signal current received at the signal track. This reduction in signal current will give rise to a signal contrast between ridges and valleys that may be visualized e.g. as a "greyscale" fingerprint image when the signals are amplified and digitized. Figure 2 shows a situation where several sensor elements 5a-d are capacitively coupled to the same signal track 10 in the described manner. Each of the sensor elements may be activated by its associated electrode, labelled 11 a-d. As can be seen from figure 2, if an activation electrode (e.g 11 a) is left to float or held at a fixed potential, there will be practically no signal current to the signal track 10 from the corresponding sensor element 5a. This lack of signal current is regardless of whether there is a finger ridge or valley above the element. This means that only those sensor pads which are "activated" by an AC voltage will give rise to signal current. This assures an effective "multiplexing" between the different subgroups using the activation electrode voltage as "control signal".

One important function of the reference plane 7 is to shield the signal tracks 10 from parasitic coupling of AC signal from the finger 1. However, with the suggested concept the coupling of AC voltage from the activation electrodes to the finger through the sensor element pad is very limited. For this reason it may only be necessary to shield the signal tracks by covering them with a sufficiently thick layer of dielectric. This may reduce the number of required metal layers on the top surface and thus the cost of the device.

In an alternative embodiment the electrode 7 may be powered with an AC voltage of inverted polarity to increase the sensitivity of the device thus providing varying reference potential.

In addition to the described version above, the invention may take the form of a number of different embodiments. For instance, the use of a "current sensing", inverting amplifier coupling is only exemplary. Other principles, for instance based on a non- inverting amplifier coupling, are possible. The invention is thus not limited to the coupling described in fig 1, and other schematics, including voltage sensing techniques may be used to measure the magnitude of the variable capacitor 14, which is the main parameter of interest.

More, the activation electrodes may apply a switched DC voltage instead of an AC voltage, thus making a DC capacitive sensing principle. In this case the amplifiers may for instance be replaced or supplemented by a combination of a counter and a voltage level detector (comparator). There may also be used an AC voltage with several frequencies, and several activation electrodes, each with a different voltage signal, may be used for each sensor element. The capacitive elements in figure may also be substituted by other impedance components or networks, or in case of a semiconductor substrate "active" switches may be used.

The geometrical arrangements of electrodes is only exemplatory. There are several ways the electrodes can be arranged to achieve the described functionality. The electrodes may for instance be positioned in the same layer so that the capacitive coupling goes "sideways". Using a sensor geometry as described in WO 97/58342, which is included here by way of reference, the activation electrode, sensor pad and signal tracks may be defined in the various layers of a PCB board. In an alternative embodiment the combination of 11 and 12 may be replaced by for instance a constant amplitude AC current source.

To calibrate the response from the sensor, a calibrating capacitor can be coupled between each signal track and another AC electrode that can be turned on and off. When the electrode is turned on, the change in signal current will represent the overall variations in measuring system, e.g. the gain of the amplifier. This reading can therefore be used to normalize the output from each channel. As the calibration capacitor contributes with a constant signal level (AC current of constant amplitude) it may also be used to shift the signal level up or down. Alternatively an AC current source may be used for calibration. In other embodiments the contact between the finger and the sensor pad may also be galvanic, and a shunting capacitance or impedance (equivalent to 13) added between the pad and the input terminal of the amplifier, as described in patent application WO 03/049012, which is included here by way of reference. Also the dielectric layer 8 may be provided with upper conductors moulded in an insulating material such as plastic, for direct contact with the finger surface, e.g. for shaping the upper surface of the sensor in order to improve the contact with the finger or to provide a sensor element structure being different from the structure being provided by the locations of the sensors 5. This provides a sensor surface which keeps its electrical characteristics even if the surface is worn down. If a laminate or other flexible type of substrate is used, the proposed principle will be very well suited for integrating the sensor with a smartcard, due to the very small size of the ASIC. Wet fingers may in many cases be a problem if the fingerprint valleys are filled with sweat or other conductive material. By adding a "local" electrode close to each subgroup of sensor elements and feeding this electrode e.g. with the same signal as the associated activation electrode, thus employing the principles described in WO 03/049011, which is included here by way of reference, the sensor may suppress the effects of very wet fingers.

If the external electrode 7 is grounded, this will make use of additional ESD protection circuitry less necessary as discharges from the finger will go directly to ground. The use of a grounded top layer may also be an advantage with respect to electromagnetic radiation to the surroundings. The described embodiment has the advantage that all active switches can be located inside the ASIC, that the substrate process is very simple, and that the number of connections between the substrate and the ASIC is very low. In addition, the surface of the sensor has no active drive circuitry which stimulates the finger with an AC signal.

Figures 10-13 illustrate an alternative embodiment of the invention. As before one column or set of sensor elements is coupled to one common amplifier. However, instead of using an AC drive electrode to activate a row of sensor elements, the control electrode is used to control a number of switches 27, which in turn are used to turn the signal from the sensor elements on and off in a predetermined sequence, so that only one row, i.e. one element in each column, is active at a time.

Said switches are preferably active semiconductor switches, and may be perferably be realized in the substrate, or alternatively on the ASIC itself. Preferably the sensor elements are fabricated on a substrate, but alternatively they may also be fabricated directly in the ASIC itself.

In case of active switches on the substrate, the substrate is preferably of silicon (or other semiconducting) or a passive substrate with integrated semiconductor switches, for instance in the form of thin film transistors (TFTs). Essentially, all other active devices (amplifiers, filters, digital circuitry) are fabricated in the ASIC itself, so that the substrate process becomes as simple as possible. A typical substrate process may be a silicon substrate with only 1-3 layers of doped semiconductor and 2-3 metal layers, which will be enough to fabricate e.g. a simple MOS switch. To save costs, the switching means should be made with as few active layers as possible. For instance, a switch can be made either as a MOSFET, JFET or a bipolar transistor controlled by the voltage on the control electrode. Such components can be made in a very simple silicon or TFT process employing only 1-2 doping layers, much less than what is needed for a full CMOS process or similar. The cost of processing can thus be held at a minimum Alternatively, a switch may be realized as a diode with a DC bias determined by the control electrode. As will be familiar to a person skilled in the art, the DC bias level can be used to control the capacitance and impedance of the diode, thus making it possible to modify the flow of current through it. The switches may be designed in such a way that their impedance can be controlled by use of the control voltage. For instance, the control voltage may be used to switch between a state of high (switch "closed") or low (switch "open") impedance. To make the principle more selective, two switches can be made in parallel: One switch that opens and closes the connection to the signal track, and one second switch that simultaneously closes and opens a connection to a reference potential, e.g ground. The point here is that when the first switch is closed and the second switch is open, the signal current will flow to said reference potential rather than to the signal track. In an alternative embodiment there is a controllable switch only to said reference potential, and the connection to the signal track is made through a fixed impedance (e.g. a capacitor).

In a preferred embodiment, a drive electrode (e.g. AC or switched DC) with a reference potential common to all elements in the essentially linear array is located on the side of the sensor elements. The amplifiers and detection circuitry is set up to measure the impedance between the drive electrode and the individual sensor elements through the finger. This may for instance be done by measuring the current that flows from the drive electrode through the fmger and the dielectric to the sensor element.

In one preferred embodiment, detailed in figure 10-13, the invention consists of a substrate 2 which may be fabricated from e.g. silicon. On the top surface of the substrate, over which the finger to be imaged 1 shall be pulled, a number of sensor elements 5 are defined, e.g. as conductive pads made in a thin film process. The sensor elements are coupled to the signal tracks and/or an optional reference potential through a switch 27 fabricated in the semiconductor substrate. The sensor elements are organized groupwise in an essentially linear array configuration. Preferably the sensor is equipped with a number signal tracks 10 leading to the input channels of the ASIC 4. To each signal track, a number of sensor elements 5 (one column 23) are coupled through a number of switches 27, e.g. one switch for each sensor element. The switches can be turned on and off by applying a control voltage on a number of activation electrodes 11. Each activation electrode 11 is coupled to one row 22 of sensor elements. When only one of the rows are turned on or activated, the signal tracks will receive a signal current from only this row only. I.e., by switching between the various activation electrodes the sensor elements can be addressed individually, one row at a time.

Said activation electrodes 11 and tracks 10 may for instance be fabricated in a conductive layer (metal or doped semiconductor) on the top side of the sensor, e.g. in the same layer as the sensor elements. To route the connections for the various columns and rows to the individual sensor elements, two conductive layers with inter-layer vias may be used. Preferably, both these layers are fabricated on the top side of the sensor substrate 2. In addition, there should preferably be one shielding layer 24 between the tracks and the finger.

As visualized in figure 13, the switch may for instance be fabricated as a MOSFET device. In this configuration, the activation or control electrode 11 serves as the gate electrode of the transistor, while the sensor element 5 and the signal track 10 are coupled to the two doped source and drain regions 25. When an appropriate voltage is applied to the gate, a conductive channel will be opened between the source and gate, and the device will be in an "open" or "on" state, as will be familiar to anybody skilled in the art. Likewise, another voltage may be used to close the channel and turn the transistor off. Extra doping layers or dielectric structures may be used to isolate the various switches from one another.

The ASIC 4 (or an equivalent electronical signal conditioning system) contains a number of amplifiers 16 to amplify the signal associated with each signal channel, as well as other signal conditioning circuitry. A schematic version of a preferred measuring principle is shown in figure 10. An AC voltage signal from the drive electrode 7 couples to the sensor element 5 through the finger and the capacitance 13 through the sensor dielectric. In case of a fingerprint ridge in contact with the surface directly above a sensor element, the variable capacitor 14 will be high. When the switch 27 is open and this element thus is activated, this will lead to a flow of AC current to the input channel and through the feedback impedance 19, giving rise to an output signal from the amplifier 16. In case of a valley, the capacitor 14 will be very small and there will be practically no flow of signal current. In case the switch 27 is closed, there will be no current flow and hence no signal regardless of the condition of the finger. The signal current to each channel can be filtered, amplified and demodulated (e.g synchronously) by appropriate circuitry in the ASIC.

Finally, the ASIC contains at least one logical circuit for feeding the activation electrodes 11 with appropriate control voltages. The activation circuit may for instance be a switchable DC potential coupled to n output channels through a multiplexer, or another switching device that feeds one activation electrode at the time with the appropriate control voltages.

Claims

C l a i m s

1. Sensor module for measuring structures in a surface, especially a finger surface, comprising number of sensor elements being located in chosen positions in a common surface, the sensor elements being arranged in an essentially linear array and an external reference potential, the sensor elements also being coupled to an electronic circuit, said electronic circuit being adapted to measure the magnitude of the capacitance or impedance between chosen sensor elements and said external reference potential at chosen points of time, said electronic circuitry comprising input means including at least two amplifier circuits, each amplifier circuit being coupled to a set of sensor elements including at least two elements for amplifying signals therefrom and transmitting it to said electronic circuitry, each sensor element in each set being coupled to a control means for activating each sensor element in a set according to a predetermined sequence within a chosen time frame, the amplifier circuit thus receiving a signal from one sensor element in a set at the time, said electronic circuit being adapted to combine the values measured at the sensor elements within said time frame to form an essentially linear image segment of the surface based on knowledge about the lαiown positions and activation times of each sensor element.

2. Sensor module according to claim 1, wherein the essentially linear array is adapted to measure a finger surface being moved over the sensor elements, said linear array also being provided with movement measuring means for measuring the movement between the sensor and the surface, and the electronic circuit being adapted to combine said linear image segments of the surface to generate a two dimensional representation of the surface structures.

3. Sensor module according to claim 2, wherein the sensor elements is positioned on a substrate, said substrate comprising conductor leads coupling the sensor elements to an electronic circuit.

4. Sensor module according to claim 3, wherein the substrate is provided with a number of openings through which openings electrical contacts being coupled to the sensor elements are lead, and that the electronic circuit is positioned on the opposite side of the substrate relative to the sensor elements.

5. Sensor module according to claim 3 or 4, wherein the control means for activating a sensor element is a drive electrode being coupled to said sensor element through an impedance, said drive means providing an excitation current or voltage to said element, thus coupling a signal through the sensor element to said external reference potential, and where said sensor element is coupled to said conductor lead through an impedance.

6. Sensor module according to claim 1 wherein each drive means is coupled to a number of sensor elements, one in each set, so as to activate one sensor element in each sensor element set simultaneously.

7. Sensor module according to claim 5, wherein the reference potential is a variable voltage.

8. Sensor module according to claim 5 or 7, wherein said external reference potential is provided by an outer electrode position apart from the sensor elements adapted for coupling said potential to said surface.

9. Sensor module according to claim 3, where the substrate is fabricated from silicon, glass, ceramics or a polymeric material.

10. Sensor module according to claim 5, 7 or 8, wherein the sensor elements are provided with a dielectric layer providing a capacitance between the sensor element and the surface to be measured being significantly larger than the capacitance between the reference potential and the surface.

11. Sensor module according to claim 2,3 or 4 where said control means comprises an electronic switch

12. Sensor module according to claim 11, where the switch is integrated in the substrate.

13. Sensor module according to claim 12, wherein the substrate is a semiconductor, and where the switches are defined in the top surface of the substrate.

14. Sensor module according to claim 11, where said switches and sensor elements are integrated in an electronic circuit

15. Sensor module according to claim 11, where said sensor elements and switches are integrated in said electronic circuit

16. Sensor module according to any preceding claim, where the sensor elements in a set form a group of essentially neighbouring elements in said essentially linear array.

17. Sensor module according to claim 16, where said predetermined sequence is so that two neighbouring elements are always accessed directly after one another.