BACKGROUND ART Keypads are widely used for entering numeric codes etc. , and typically consist of a unit with a number of accessible push buttons. The most common design is a twelve button keypad, and this is usually arranged in a rectangular 3x4 array with the digits 1-9 in the upper three lines, together with three additional push buttons forming the bottom line. The central button of the bottom line is designated as the zero key, and the two buttons either side are used for additional functions and typically marked * and # or the like. Other combinations are however possible, together with other arrangements and layouts.

One use of keypads is for use in access control, for example by requiring users to enter a numeric code in order to release a lock and gain access to a controlled area. In this case, the keypads may have to be accessible from the outside of the controlled area or building and this presents significant difficulties.

Push buttons require moving parts and are subject to wear and other similar degradation processes, and exposing the unit to the external environment can introduce additional challenges in ensuring its reliability. One proposal is to use piezo elements underneath an effectively stationary button. The fact that the push button does not move significantly allows it to be sealed against the external environment. This offers some advantages in use, but the piezo elements are expensive and significant force is usually required to activate them.

Users with a lighter touch can thus find the unit to be unreliable.

It has also been proposed to sense a users finger optically. At the location of each push button, an infrared beam is emitted from the unit, and if a finger is placed on or over the push button then this infrared beam will be reflected into a suitably positioned photodiode or other optical sensor.

No force is required and no moving parts are involved. However, optical systems such as this have hitherto suffered from unreliability in use due to the wide variation in ambient light that can be encountered, particularly when direct sunlight impinges on the unit.

SUMMARY OF THE INVENTION The present invention therefore seeks to provide a reliable keypad that is able to alleviate such difficulties.

The present invention therefore provides a keypad comprising a light- emitting element, and a light-detecting element adapted to produce an output voltage dependent on the detected light level, and a processing means adapted to interpret the output of the light-detecting element, including a filter between the light-detecting element and the processing means adapted to block the DC element of the output voltage at least during measurement thereof by the processing means.

It will usually be preferred that the keypad includes a plurality of keys, each key including an associated light-detecting element and (usually), each key including an associated light-emitting element. The outputs of the plurality of light-detecting elements are preferably multiplexed in a multiplexing element and fed to a single processing means. In this case, the filter preferably acts on the output signal between the multiplexing element and the processing means.

The filter can include a capacitive element in series connection to block the DC signal. A resistive element in parallel connection is also preferred.

The filter preferably includes a reset line which, when activated, zeros the output thereof. A transistor arranged so as to ground the output of the filter when the reset line is activated is suited to this task. The processing means can then activate the reset line prior to taking a measurement.

It is preferred that each key is located behind a circular aperture, especially an aperture with a bevelled edge. This assists in shading the key from external sunlight when a finger is placed over it.

In a preferred form of the keypad, the processing means interprets the output of the light-detecting element a plurality of times and records a keypress if the output is consistent therewith on all of those times. This inhibits false positives caused by transient events, such as a reflection from a passing vehicle or the like. The processing means can suitably interpret the output of the light- detecting element between two and five times, such as three or four times. It is then still more preferable to adjust the frequency of the repeated scans such that the processing means will not be falsely triggered by strong AC mains lighting. This is powered at 50 or 60Hz and thus exhibits a stroboscopic effect at 100 or 120Hz. Thus, it is advisable that the time interval between the successive scans not a simple multiple of half the period of AC mains power supply.

A suitable light-detecting element is a phototransistor, and a suitable light-emitting element is a light emitting diode (LED). These preferably operate in the infra-red (IR) region, but other wavelengths such as red, green, blue etc are also usable.

BRIEF DESCRIPTION OF THE DRAWINGS An embodiment of the present invention will now be described by way of example, with reference to the accompanying figures in which; Figure 1 shows a keypad according to the present invention; Figure 2 shows a section through a single key according to the present invention with a finger in place operating the key; Figure 3 shows a single key without a finger present; and Figure 4 shows the filter circuitry according to the present invention.

DETAILED DESCRIPTION OF THE EMBODIMENTS Figure 1 shows a keypad according to the present invention. A case 100 has twelve apertures 102 each covered with a transparent plate 104, and collectively defining a 4x3 array of keys as described above. The invention could of course be applied to other keypad layouts, but the illustrated layout is the most common.

A pair of feedback LED elements 106,108 are also provided in the front face of the case 100. These illuminate green or red to indicate success or otherwise in entering a code.

Figure 2 shows the arrangement of a single key. An infra-red (IR) LED 110 is mounted on a printed circuit board 112 adjacent an IR-sensitive phototransistor 114. Other forms of optoelectronic devices with similar functions could of course be employed. A moulded plastic wall is held in place between the LED and the phototransistor to limit direct transmission between them. As shown in figure 2, when a finger 116 is placed above the key, IR transmissions from the LED 110 are reflected back to the phototransistor 112.

Figure 3 shows an alternative transmission path that operates even in the absence of a finger. IR emissions by the LED 110 can be coupled into the transparent plate 104 and reflected internally. Some of this light will exit and be captured by the phototransistor 114. Thus, there will always be a certain level of response from the phototransistor 114 when the LED 110 is activated, the precise amount depending on the manufacturer's performance tolerances of the phototransistor and LED pair used at a particular key and the reflectance of the surface above the key. During production test of the keypad assembly the processing means measures, for each key, the incremental levels of light detected by each phototransistor 114 when its associated LED 11 is suddenly turned on with no finger present. The processing means then calculates and stores a threshold level above which it later assumes a finger must be present above a key to cause that increased level of reflected light to be detected. Such an increased level is of course in addition to the ambient light level that is detected continually by the phototransistor and must be distinguished from signals caused by the user input.

Many existing designs of IR keypads attempt to use a"store and hold" arrangement to do so. In such systems, the phototransistor output is checked and stored from time to time, and the instantaneous signal is compared to the last stored signal to detect increases above and beyond the stored signal. Such increases, if sufficient in magnitude, indicate the application of a finger to the key. This approach has a significant problem in that direct sunlight impinging on the keypad 100 will often saturate the phototransistor 114. This means that a system looking for a rise in the phototransistor output relative to the recent historical mean signal will be disabled by direct sunlight.

Figure 4 shows the detection circuit according to the present invention. It detects a change in light level seen by a phototransistor 114 when an adjacent IR LED 110 is suddenly turned on, even when the keypad 100 is subject to very high levels of sunlight or bright AC lighting. When a finger is placed on a key the IR light level reflected back to the phototransistor increases by 3 to 4 times, compared to the level when no finger is present.

At each key, the phototransistor 114 allows a variable current to drain depending on its incident light level. The associated LED 110 is grounded via a transistor 118 controlled by logic circuitry in the remainder of the keypad so that each key is activated as required. This is duplicated the required number of times (for example twelve), of which two such instances are shown in figure 4.

A multiplexer 120 is addressed to channel the output current from one phototransistor 114a, 114b etc to the sensor circuitry. The signal is output essentially unchanged as output line 122, grounded via a resistor 124. The output line 122 is also connected to a capacitor 126 in series, and selectively grounded via an NFET transistor 128. That selectively grounded output is fed to the remainder of the logic circuitry as the output of the selected key.

In practice, the phototransistor may be in the dark or at very high light level, and the resulting wide variation of photocurrent can generate a wide range of voltages. A processor holds the IR LEDs off initially (via transistors 118) and holds the transistor 128 on. Thus, with one side of capacitor 126 held at ground by transistor 128, the capacitor 126 quickly charges up to the voltage on resistor 124 that is generated by the photocurrent from the key being addressed. After a short delay to let the circuitry settle, the processor releases transistor 128 and also turns on an IR LED via transistor 118a or 118b etc.

When an IR LED turns on, the IR level at the key suddenly increases and the increased level is sensed by the phototransistor. The AC voltage change at the resistor 124 caused by turning on the IR LED is coupled via capacitor 126 to the output at 130. After a delay of about 100, us, to allow the IR photocurrent to stabilise, the changed value is read by an analogue to digital converter (A/D) and compared to a stored threshold for that key. All twelve keys are scanned in turn. Keys are scanned continually until one is seen to have a level above its threshold, when an output is given to indicate that the key has been pressed.

The transparent plastic cover against which the Opto devices press does reflect some IR light back from the IR LED to phototransistor, but it is a relatively low level and increases by 3-4 times when a key is pressed. Most fingers give a good IR response-reduced a little in damp conditions. Many glove materials also provide a good IR reflection.

In full sunlight the phototransistor outputs can saturate, raising resistor 124 to the maximum (supply) voltage. However, when a finger presses a key it shades it from the sunlight putting the resistor 124 back into an active region where a change of IR level is readily seen as the LED turns on. The aperture around the key is of circular shape, with a bevelled profile, of size and shape to promote the shading effect of a fingertip placed on a key. However, sunlight can be of such intensity that light passing through a fingertip can still present a very high ambient light level at a shaded key. The illustrated circuit provides means to detect the increased level of light when an LED is pulsed over a very wide range of ambient light levels.

In practice, the processing means scans the keypad repeatedly and swiftly. A key is regarded as having been"pressed"if a finger is detected during three successive scans. The signal level indicating the presence or absence of a finger must thus be detected 3 times in succession by the processor before it is registered as a change of state. This avoids detection of single spurious background light variations, for example light reflected from the windscreen of a passing car, or a modulated light level generated by AC powered lighting. A level above the"finger present"threshold must be seen 3 times in succession to accept that a finger is there and must not be seen 3 times in succession to accept it is gone. The key scan rate is therefore chosen so that the 3 successive readings do not coincide with the successive maximum rate of change of light levels that are generated by 50 or 60Hz AC powered lighting. This makes the keypad more immune to high levels of AC powered lighting. Of course, the processing means could scan the keypad more or less than three times if desired.

Each key On/Off state is independently detected. This helps because a spurious key-press may be sensed when condensation gathers on a key surface but will be cleared when a finger presses that key and disperses the condensation, so allowing that key to be detected normally again. If the keypad stopped on a particular key when it was sensed, and did not continue to look at all the other keys, a finger would not be detected if it pressed another key and cleared the condensation at that key.

It will of course be understood that many variations may be made to the above-described embodiment without departing from the scope of the present invention.