WO2007020576A1

WO2007020576A1 - Active matrix display devices

Info

Publication number: WO2007020576A1
Application number: PCT/IB2006/052769
Authority: WO
Inventors: Steven C. Deane; Ian D. French; Alan G. Knapp; Paul Collins; David A. Fish
Original assignee: Koninklijke Philips Electronics NV
Current assignee: Koninklijke Philips NV
Priority date: 2005-08-16
Filing date: 2006-08-10
Publication date: 2007-02-22
Anticipated expiration: 2008-02-16

Abstract

A display device comprises an array of rows and columns of pixels disposed over a common substrate. Each pixel comprises a display pixel element having a stable drive state in which the optical characteristics remain substantially constant and a drive circuit for controlling the voltage applied to the display pixel element. An optical sensor arrangement of each pixel (PR1, R2) detects an optical signal dependent on the setting of the display pixel element, and the drive circuit comprises an optical feedback control system which is arranged to drive the display pixel element to the stable display state in response to an applied input pixel voltage and an output of the optical sensor arrangement, the optical characteristics being varied before the drive of the display pixel element reaches the stable state and being held substantially constant once the stable state has been reached. This approach removes the need to apply complicated voltage control schemes to the display element. Instead, a desired brightness is loaded into the pixel (as a voltage level), and optical feedback within the pixel is used to achieve the desired brightness level.

Description

DESCRIPTION

ACTIVE MATRIX DISPLAY DEVICES

This invention relates to active matrix display devices, in particular electrophoretic active matrix display devices.

Electrophoretic display devices are one example of bistable display technology, which use the movement of particles within an electric field to provide a selective light scattering or absorption function. In one example, white particles are suspended in an absorptive liquid, and the electric field can be used to bring the particles to the surface of the device. In this position, they may perform a light scattering function, so that the display appears white. Movement away from the top surface enables the colour of the liquid to be seen, for example black. In another example, there may be two types of particle, for example black negatively charged particles and white positively charged particles, suspended in a transparent fluid. There are a number of different possible configurations.

It has been recognised that electrophoretic display devices enable low power consumption as a result of their bistability (an image is retained with no voltage applied), and they can enable thin display devices to be formed as there is no need for a backlight or polariser. They may also be made from plastic materials, and there is also the possibility of low cost reel-to-reel processing in the manufacture of such displays.

For example, the incorporation of an electrophoretic display device into a smart card has been proposed, taking advantage of the thin and intrinsically flexible nature of a plastic substrate, as well the low power consumption.

In the simplest form, a passive matrix addressing scheme is used. Figure 1 shows a known passive matrix display layout for generating perpendicular electric fields between the top column electrodes 10 and the bottom row electrodes 12. The electrodes are generally situated on two separate substrates. The passive matrix electrophoretic display comprises an array of electrophoretic cells arranged in rows and columns and sandwiched between the top and bottom electrode layers. The column electrodes 10 are transparent. Cross bias is a problem in the design of passive matrix displays. Cross bias refers to the bias voltages applied to electrodes that are associated with display cells that are not in the scanning row (the row being updated with display data). For example, to change the state of cells in a scanning row in a typical display, bias voltages might be applied to column electrodes in the top electrode layer for those cells to be changed, or to hold cells in their initial state. Such column electrodes are associated with all of the display cells in their column, including the many cells not located in the scanning row.

A further problem associated with the use of passive matrix addressing is that the driving signals must be introduced to the display sequentially, typically one line at a time, along the (orthogonal) selection rows and data columns. Once the line is no longer being addressed, the electrical field is reduced to a level whereby the particles will not move. As a consequence, the particles only move whilst a line is addressed, and it will take a long time to complete addressing the display (in general, the response speed of the pixel times the number of rows in the display). As the display operates using the physical movement of particles, there is a limit to the speed at which a pixel can be addressed.

In order to speed up the addressing and to overcome the cross bias problem, it is known to use active matrix addressing, which ensures that the driving voltage is maintained during the time that other lines of the display are being selected, and also provides electrical isolation of pixels from the signal lines when not being addressed.

In an active matrix display, switching elements such as diodes or transistors are used, either alone or in conjunction with other elements, to control pixel electrodes associated with the display cell or cells associated with an individual pixel. In one typical active matrix display configuration, for example, a common potential (e.g., ground potential) may be applied to a common electrode in the top layer and pixel electrodes located in the bottom layer are controlled by associated switching elements to either apply a biasing voltage to the pixel electrode or to isolate the pixel electrode to prevent an electric field from being generated that would cause the associated display cell(s) to change state.

Electrophoretic display devices can use the movement of particles in a number of ways. In a system generating vertical electric fields, as shown in Figure 1 , the particles are controlled to move selectively up and down the display material layer. When the particles are at the top, they are visible, and when they are at the bottom, then they are not visible, and the medium supporting the particles is then visible. The particles may be white, and the supporting medium may be red, green or blue. The term "transverse electric field" is used below to refer to this vertical electric filed, perpendicular to the planes of the substrates.

Another type of electrophoretic display device uses so-called "in plane switching". This type of device uses movement of the particles selectively laterally in the display material layer. When the particles are moved towards lateral electrodes, an opening appears between the particles, through which an underlying surface can be seen. When the particles are randomly dispersed, they block the passage of light to the underlying surface and the particle colour is seen. The particles may be coloured and the underlying surface black or white, or else the particles can be black or white, and the underlying surface coloured.

An advantage of in-plane switching is that the device can be adapted for transmissive, reflective or transflective operation. In particular, the movement of the particles creates a passageway for light, so that both reflective and transmissive operation can be implemented through the material. Monochrome electrophoretic display systems are used for electronic reading devices, whilst colour versions are being developed for signage and billboard display applications, and as (pixellated) light sources in electronic window and ambient lighting applications.

A monochrome electrophoretic display typically uses black and/or white particles within a transparent fluid. A number of ways are being explored to implement a colour display. For reflective colour displays, the use of colour filters is not attractive, as there is an associated loss of brightness. For transmissive/transflective displays, the use of a backlight makes colour filtering more appropriate.

One approach is therefore to provide a white backlight, and to use colour filtering to convert a monochrome output into a colour sub-pixellated display. One example of the use of a colour filter is disclosed in WO

04/074921.

An alternative is to provide a backlight which flashes in three different colours, and to control each colour output in sequence. This may use a monochrome pixel array, essentially functioning as a light valve for each colour, although this requires higher speed operation which may not be suitable for existing electrophoretic display technology.

An alternative is to use different colour particles to implement a colour filtering operation within the pixel. For example, the black level can be made in a subtractive way by absorbing red, green and blue parts of the backlight spectrum by moving cyan, magenta and yellow electrophoretic particles in a transparent fluid into the light path. White is made by moving all of these coloured particles out of the light path into a so-called "container".

This approach enables a white backlight to be used, and the pixel output colour is obtained with one addressing phase. This approach does however require three different types of particle which can be moved independently between the container and the pixel aperture. This can be achieved by having particles which move with different speeds, and using these differences to devise a control scheme which enables selected particles to be moved to the pixel aperture. Such an approach is described in WO

2004/088409 and WO 04/066023. Different frequency responses of the particles has also been proposed as a way of providing independent driving of each colour particle.

Electrophoretic displays are driven by complex driving signals, particularly if grey scales are to be enabled. For a particle to be switched from one grey level to another, often it is first switched to white or black and to then to the final grey level. This can lead to visible artefacts of the image during transition, in particular a highly undesirable flashing of the image can occur.

These complex schemes are needed because, unlike LCDs, electrophoretic displays respond to the integral of voltage with time (rather than the RMS voltage), so errors accumulate. In addition, the response has dependencies on previous images, temperature, etc, that need to be compensated for.

These drive schemes need complex drive electronics, including several frame stores and non-standard column driver chips (typically with +/-15V range). The complex drive schemes require significant time and power for a display refresh, and even then can only achieve a limited number of greyscales.

This invention relates to all of the above types of transmissive or transflective colour or monochrome displays using active matrix addressing, and relates to simplified addressing schemes for providing greyscale operation.

According to the invention, there is provided a display device, comprising an array of rows and columns of pixels disposed over a common substrate, wherein each pixel comprises: a display pixel element having a stable drive state in which the optical characteristics remain substantially constant; a drive circuit for controlling the voltage applied to the display pixel element; an optical sensor arrangement for detecting an optical signal dependent on the setting of the display pixel element; wherein the drive circuit comprises an optical feedback control system which is arranged to drive the display pixel element to the stable display state in response to an applied input pixel voltage and an output of the optical sensor arrangement, the optical characteristics being varied before the drive of the display pixel element reaches the stable state and being held substantially constant once the stable state has been reached. This display uses a driving method which involves varying the optical characteristics of the display device pixels until an optical feedback system indicates that a desired optical output has been reached. This approach removes the need to apply complicated voltage control schemes to the display element. Instead, a desired brightness is loaded into the pixel (as a voltage level), and optical feedback within the pixel is used to achieve the desired brightness level. This feedback system will only reach the stable state when the desired optical response has been reached, and therefore avoids the need to take into account previous display history when determining the voltage to be applied to the pixel. Simple drive circuitry can therefore be used which applies voltages based on a single mapping function of light output level to pixel drive voltage. The optical feedback control system thus implements the corresponding function between the optical output of the display pixel element and the applied input pixel voltage.

The optical feedback control system can be arranged to drive the display pixel element to a condition in which OV is applied across the display element.

The drive circuit can comprise an input capacitor for charging to the input pixel voltage. This can be connected at one terminal to a data input line though an address transistor. This provides a simple pixel addressing scheme, which involves controlling the address transistors to load a pixel voltage onto a storage capacitor, and this is the same technique used in other display technologies.

The optical sensor arrangement may comprises a light-sensitive potentiometer, and the input capacitor is then connected at its other terminal to the potentiometer output. Thus, the capacitor and the potentiometer are effectively in series and together form the feedback control loop. The drive circuit can comprise an inverter having as input a signal which is dependent on the output of the optical sensor arrangement and the applied input pixel voltage. This inverter operates to change the drive voltage applied to the display element to move the optical output to the desired level. The drive circuit may instead comprise a reset portion for driving the display pixel element to a reset state, and a drive transistor for changing the optical state until the stable display state is reached. This design changes the optical state in one sense only (for example from dark to light or vice versa) from a reset state to the desired state. A pulsed output transistor can be provided between the output of the optical feedback control system and the display pixel element, and this can provide power savings.

In another version of the drive circuit, a drive transistor is for changing the optical state of the display pixel element independently of the applied input pixel voltage, and a reset portion drives the display pixel element to the stable display state at a time dependent on the applied input pixel voltage and an output of the optical sensor. This drive circuit design can change the optical state using a constant drive voltage, and halt the change more quickly, in a snap-off manner. The reset portion may comprise a transistor in parallel with the display pixel element, or it may comprise a plurality of cascaded transistors, one of which is in parallel with the display pixel element. The latter approach provides a faster snap-off response.

An illumination source may be required for providing a reference illumination to the optical sensor arrangement, particularly for reflective displays. The optical sensor arrangement can then comprise first and second optical sensing elements, one of which is responsive to the illumination source only and the other of which is responsive to the display pixel setting.

The optical sensor arrangement can be arranged to detect a light intensity of light reflected from the opposite side of the display pixel to the output side of the display pixel, the reflected light being reflected light of the illumination source. The invention is of particular applicability to an electrophoretic active matrix display device.

The invention also provides a method of driving a display device, the display device comprising an array of rows and columns of pixels disposed over a common substrate, wherein each pixel comprises a display pixel element having a stable drive state in which the optical characteristics remain substantially constant, wherein the method comprises, for each display pixel: varying the optical characteristics by applying a voltage to the display pixel element; using an optical feedback signal dependent on the setting of the display pixel element and dependent on an applied input pixel voltage to drive the display pixel element to the stable display state, and using in-pixel optical feedback and drive circuitry; and holding the optical characteristics substantially constant once the stable state has been reached.

Examples of the invention will now be described in detail with reference to the accompanying drawings, in which:

Figure 1 shows a known passive matrix display layout; Figure 2 shows a first pixel design of the invention;

Figure 3 shows alternative inverter designs for use in the pixel design of Figure 2;

Figure 4 is used to explain the operation of the pixel design of Figure 2;

Figure 5 shows a second pixel design of the invention; Figure 6 is used to explain the operation of the pixel design of Figure 5;

Figure 7 is used to compare the response of the circuits of Figures 2 and 5;

Figure 8 shows a third pixel design of the invention;

Figure 9 is used to explain the operation of the pixel design of Figure 8; Figure 10 shows a fourth pixel design of the invention;

Figure 11 is used to explain the operation of the pixel design of Figure 10; Figure 12 shows an in-plane pixel design to which the invention can be applied;

Figure 13 shows a fifth pixel design of the invention; and Figure 14 shows a sixth pixel design of the invention.

The same references are used in different Figures to denote the same components, and description is not repeated.

The invention relates to display device technologies in which display pixels have a stable drive state in which the optical characteristics remain substantially constant. One example discussed above is electrophoretic displays, in which the movement of particles is substantially halted once an electric field is removed, but there are other examples of display technology of this type, such as electrochromic displays. Electrochromic displays are held in a stable state when there is no current flow, but a voltage is applied. The invention provides a pixel drive circuit for an active matrix pixel array for this type of display in which the optical setting of the display pixel element is detected and used as a feedback parameter, together with a pixel setting provided as input. The display pixel element is driven to the stable display state using this feedback, with the optical characteristics being varied before the drive of the display pixel element reaches the stable state, so as to stabilize the feedback loop.

Simple external drive circuitry can be used which applies pixel drive voltages based on a single mapping function of light output level to pixel drive voltage. The in-pixel optical feedback system then implements the required function between the optical output of the display pixel element and the applied input pixel voltage.

Figure 2 shows a first example of pixel circuit of the invention. The circuit comprises an input capacitor C1 for charging to the input pixel voltage which is provided on a column data line. The input capacitor C1 is connected at one terminal, Node 1 , to the data input line though an address transistor T1. An optical sensor arrangement is in the form of a photoresistor PR1 and a resistor R2 in series between power lines VP1 and VP2. The photoresistor PR1 detects the setting of the pixel, by sensing light scattered back from the display pixel. This arrangement acts as a light-sensitive potentiometer, and the input capacitor (C1) is connected at its other terminal to the potentiometer output, Node 2. The voltage of this node varies in dependence on the setting of the display element 20, represented as a capacitance. This capacitance 20 is the pixel and associated storage capacitance.

A drive circuit provides the output to the pixel 20 and comprises an inverter T3.T4, having as input, Node 1 , one terminal of the capacitor C1. The capacitor C1 is effectively in series with the light sensitive potentiometer, so that a feedback path is defined which includes both the pixel drive voltage (stored on C1 ) and the optical feedback element. Thus, the input to the inverter is dependent on the output of the optical sensor arrangement and the applied input pixel voltage.

A loading transistor T2 is provided for enabling the pixel drive voltage to be stored on the capacitor C1 , and this is connected between the second terminal of the capacitor and the low voltage rail, effectively bypassing the resistor R2 when data is loaded onto the capacitor C1.

As shown, a row line controls the address transistor T1 and the loading transistor T2 for the step of loading pixel data into the capacitor C1. Once the data is loaded, the two transistors are turned off, and the charge is held on the capacitor. The charge then remains almost constant, as the terminal of the capacitor which defines Node 1 is only connected to a TFT gate.

The circuit can be arranged to fit in a typical pixel, and can be made in standard (low cost) amorphous silicon active matrix technology, although low temperature polysilicon versions are of course also possible.

A useful feature of present electrophoretic reflective displays is that they show the inverse image on the rear side. By using this inverse image as the control, the circuitry can be placed behind the display, so that nothing is visible to the end user, which could diminish the display quality. Alternative designs for the inverter stage are possible, and Figure 3 shows two examples. The design on the left is a CMOS inverter, which can be used if the active matrix is made in low temperature polysilicon (LTPS) technology, which has p-channel TFTs available. The design on the right is suitable for an amorphous silicon (a-Si) active matrix and uses a resistor Rload as the load for the inverter. Since making such resistors can be difficult in standard technology, this load can be a photo-sensitive resistor made from a- Si whose value is set to the desired level by the illumination from the back of the display which is used for sensing the pixel state.

The operation of the pixel will now be described, with reference to Figure 4. In the following description it is assumed that one side of the electrophoretic pixel is connected to ground (i.e OV) and that the top power rail VP1 is positive and the bottom power rail VP2 is negative, so that the inverter circuit can cause the voltage across the electrophoretic material in the pixel to be positive, negative or zero. It is assumed that a negative voltage causes the back of the pixel (the part seen by the sensors in the pixel circuit) to become darker which means that a viewer would see the pixel becoming brighter. A positive voltage has the opposite effect and a zero voltage will hold the pixel in its current state, and this is the stable state of the display pixel.

The sensing element is shown as using a photo-resistor, PR1 , although other photo-sensing devices such as photo-sensitive transistors can also be used. This senses the light level scattered back from the rear of the electrophoretic layer. In the simplest embodiment of the circuit, the second resistor, R2, is a simple resistor that is not photosensitive. PR1 and R2 are connected in series between the positive and negative power rails defining the potentiometer.

If the light level produced by the rear illumination source is L and the reflectance of the electrophoretic layer is R_en then PR1 receives an illumination level kLR_efi where k is a constant related to the detailed structure of the pixel. Variation in the scattered light level alters the relative value of the resistance of PR1 and hence determines the voltage, V2, at their junction (Node 2 of the circuit) as V2 is given by:

V2 = VP2 + (VP1 - VP2). R(R2)/(R(PR1 ) + R(R2)) R(R2) is the resistance of the resistor R2, R(PRI ) is the resistance of the photosensitive resistor PR1 , and VP1 and VP2 are the power rail voltages. Since the resistance of the photoresistor, R(PRI), is dependent on the reflectance, R_en, of the electrophoretic pixel, the voltage, V2, is a measure of the state of the pixel and can be used as the sensing element in a feedback loop which allows the pixel to be driven to a pre-determined state defined by an input signal loaded onto capacitor C1.

In a preferred embodiment, resistor (R2) is also a photo-resistor and is arranged to sense only the light directly from the rear illumination source while the other one, PR1 senses only the light level scattered back from the rear of the electrophoretic layer. By adjusting the properties of the two photo-resistors it is then possible to arrange that the output voltage of the potentiometer V2 is largely independent of the absolute value of the illumination level used for sensing, so that this illumination level does not need to be uniform. To obtain these characteristics, the resistor (R2) can have a light blocking layer above it, and the resistor (PR2) can have a light blocking layer below it. The light source used will not alter the characteristics of the display, as it can be relatively dim, and a reflective display is relatively optically dense.

In operation, the row that is connected to the gates of the transistors T1 and T2 is pulsed high, turning both of them on. This connects one end of the storage capacitor C1 (Node 2) to the fixed voltage of the lower voltage rail, VP2, and the other end to the column. C1 is therefore charged to a voltage equal to Vdat = Vcolumn - VP2. After a time long enough for this charging to be substantially complete, the row voltage is taken low, isolating C1 from the column. During this process, Vdat will be changed slightly to a new value Vdat¹ by charge redistribution effects in the TFTs which are well known form other active matrix displays and are easily compensated by adjustment of the column voltage.

At the end of this addressing period, the voltage at Node 1 , which is the input to the inverter stage, is given by:

V1 = Vdat¹ + V2 = Vdat¹ + VP2 + (VP1 - VP2). R(R2)/(R(PR1 ) + R(R2)) This voltage will define the voltage, V3, at the inverter output, Node 3, which drives the pixel. There are three possible conditions that can apply:

V3 = 0

The pixel is in the desired state and stays in that state.

V3 is positive

This means that the pixel (as seen by the viewer of the display) is too bright. The back of the pixel is therefore too dark and the photo-resistor, PR1 , has a high value. Because V3 is positive the particles in the electrophoretic material in the pixel will move so as to make the viewed side of the pixel darker and the back side brighter. This causes the resistance of PR1 to fall, which increases V1 and reduces V3 toward OV.

V3 is negative

This is the case illustrated in the timing diagrams of Figure 4. The addressing sequence has a select period 40a and a driving period 40b.

The top plot 41 show the row address pulse which is used to charge the capacitor C1 to the data voltage. Plot 42 shows the column voltage, which is set at the pixel drive level during the row pulse. The shaded area 43 shows the time when the column is used to address other rows of pixels, and this addressing of other rows can commence at any time after the row is driven low. Plot 44 shows the inverter input voltage (Node 1 ), and this undergoes a step change when the row pulse ends. Plot 46 shows the potentiometer output (Node 2), which also has the same step at the end of the row pulse because the voltage across the capacitor C1 is constant. Subsequent changes to the voltages on Nodes 1 and 2 are the result of changes in output light intensity. It is noted that the change in gradient at point 47 is the result of a change in scale in the x-axis (time) to enable a meaningful graph to be shown. Plot 48 shows the output voltage driving the display element (Node 3). In the example shown in Figure 4, the pixel (as seen by the viewer of the display) is too dark. The back of the pixel is therefore too bright and the photo-resistor, PR1 has a low value. Because V3 is negative, the particles in the electrophoretic material in the pixel will move so as to make the viewed side of the pixel brighter and the backside darker. This causes the resistance of PR1 to increase, which decreases V1 (plot 44) and increases V3 toward OV (plot 48).

The effect of the circuit is to drive the pixel into a state of equilibrium where V3 = 0 and no further change in the state of the electrophoretic material in the pixel occurs. The state for which this equilibrium occurs depends on the voltage, Vdat¹, on C1. The more positive the Node 1 is relative to Node 2, the higher the value of the resistance of PR1 required to achieve equilibrium. In other words a more positive voltage on Node 1 relative to Node 2 produces a brighter pixel as seen by the viewer of the display. By using a simple in-pixel circuit as shown above, simple drive schemes can be used, standard column driver ICs can be used, and it is possible to speed up the display response, and reduce the display power. Indeed, improvements of a factor of ten in these two areas may be obtained compared to existing drive schemes. The pixel circuit uses optical feedback to measure the response of the display effect, and drives it until the response matches the input data. This automatically compensates for all history and temperature effects, and additionally can result in a faster image transition. In addition, far better greyscales can be achieved. Although the circuit of Figure 2 provides the required control of the pixel brightness, it has the disadvantage of consuming relatively high power. An improvement can be achieved by operating the circuit in a pulsed or sampled mode.

This approach is implemented in the circuit of Figure 5, which corresponds to Figure 2 but has an additional output transistor T5 provided between the output (Node 3) of the optical feedback control system and the display pixel element. This is controlled by a dedicated control line electrode "Sample", which can however be common to all pixels. This Sample electrode is pulsed positive at regular intervals as illustrated in the timing diagram of Figure 6. Figure 5 also shows a parasitic capacitance C2.

Figure 6 shows the output of the illumination source used to provide an optical signal for reflection by the pixel as 60, and shows the sample control line as 62. Plots 44,46 and 48 correspond to the plots of Figure 4, and in this case the output of the inverter, plot 48, is not provided continuously to the display element. Instead, the display element output is plot 64.

The rate of pulsing depends on the response speed of the electrophoretic material but would typically be every few milliseconds. Although Figure 6 shows a pulse occurring at regular intervals, it is also possible to have a variable period between successive pulses.

The illumination source used for illuminating the sensing arrangement is also pulsed in this example (plot 60) which saves significant power. The operation is similar to that described above except that the voltage at Nodes 1 and 2 will drift during the period when the illumination is off and will return to the correct value when the illumination is turned on with a speed which depends on the time constant defined by PR1 and R2 in parallel together with the stray capacitance, C2, between node 2 and ground. The capacitance C2 can thus be used to assist stable operation of the circuit during pulsed operation. The capacitances and resistors then act to fix the voltage on the nodes 1 and 2 in the non-illuminated state of the optical feedback system. The pulsed light source can avoid the need for the pulsing transistor T5.

For pulsed illumination implementations using the transistor T5, the minimum period of illumination during each sample period must be sufficient for the voltage on Node 2 to settle to the correct voltage. At some point during the period of illumination, the sample electrode is pulsed, turning T5 on and transferring the voltage at the output of the inverter (Node 3) onto the pixel and associated storage capacitance, Node 4. The transistor T5 remains on for a period that is long enough for T5 to charge the pixel capacitance to within a pre-defined error voltage that would normally be chosen to be below 10OmV. T5 is then turned off and the illumination can also be turned off. The voltage on the pixel, plot 64, now varies in discrete steps rather than continuously.

In addition to saving power, this circuit has the advantage of reaching equilibrium faster than the circuit where the voltage are applied continuously because the voltage applied to the electrophoretic material does not change continuously but remains at a constant, higher, voltage for a period so the material switches more rapidly.

This is illustrated in Figure 7, in which the top plot shows the operation of the circuit of Figure 2 and the bottom plot shows the operation of the circuit of Figure 5.

In each case, plot 70 shows the voltage at the input to the inverter (Node 1 ), plot 72 shows the voltage at the pixel, and plot 74 shows the brightness at the back of the pixel. All operating parameters the same for the two examples, and the comparison shows that in the pulsed mode, the pixel brightness reaches its final value about 25% faster.

In a further improvement, it may be desirable to reduce the voltage on the power rails VP1 and VP2 during the part or all of the period when T5 is off as this prevents power being wasted in the inverter circuits (T3 and T4) in each pixel during a period when they are not driving the pixel. The polarity of the display effect may require that PR1 and R2 are swapped to arrange for the feedback to converge (i.e. be negative). This is a simple design adjustment for a given display effect. This same technique may be applicable to other reflective displays based on other electro-optic effects such as electrochromic displays. An alternative pixel circuit and drive scheme is illustrated in Figures 8 and 9. This circuit uses a source follower instead of an inverter to drive the display pixel.

The pixel circuit has the same storage capacitor C1 and potentiometer optical feedback device. The drive circuit has a reset portion in the form of a transistor T4' for driving the display pixel element to a reset state, and a drive transistor T3' for changing the optical state until the stable display state is reached. The basic principle is to reset the pixel to a known state and then use a simple source follower circuit (which can only pull the pixel voltage in one direction) to drive the pixel to the desired state which is sensed by the photo- feedback loop. This feedback again causes the drive voltage to fall to zero when the desired state is achieved.

The reset portion, transistor T4' is driven by one extreme voltage rail (the low rail in this example) and the source follower is driven by the other voltage rail (the high rail in this example).

This circuit has the advantage that there is no standing current flow as there is in the inverter circuit, so the power consumption will be lower.

However, it does require a reset phase during which the pixels are set to black.

The operation is described below and shown in Figure 9.

Figure 9 shows a reset and load phase 90a and a drive phase 90b, and shows the row pulse as plot 91 , for charging the capacitor C1. The column voltage is plot 92.

The input (Node 1 ) to the source follower circuit is plot 94, the potentiometer output is plot 96 (Node 2) and the pixel output is plot 98 (Node 3). The signal applied to the reset line, which can control both the loading transistor T2 and the reset transistor T4', is shown as plot 99. The illumination is only present during the drive phase 90b, not during the reset and load phase 90a of operation.

A pulse is applied to the reset line, turning T4' on and pulling Node 3 (the pixel) to the negative voltage rail. At the same time, the loading transistor T2 is turned on pulling Node 2 down to the negative voltage rail as well. The pulses on the column line are the column pulses for the addressing of other rows of pixels, which at that time have their row pulse high. The addressing of all rows takes place in the phase 90a.

The fact that there is a negative voltage on the pixel means during the reset phase means that (for this example of pixel) the viewers side of the pixel will be reset white and the backside, which is sensed by the photo-sensor will go black. While the pixels are being reset, data is loaded onto the data storage capacitors by addressing via TFTs T1 and T2

At the end of the reset and load phase, the pixels are all in a known state (white to the viewer) and all pixels are loaded with their data. The reset pulse is removed and Nodes 1 ,2 and 3 all move up to their normal operating voltages defined by the values of R2 and PR1. This is seen in Figure 9. Again, the scale of the time axis is different for the two phases 90a,90b.

If the pixel is to be driven to a level different from white (viewer side) then the voltage on node will have been set in the data loading process to a value such that Node 3 is positive. This will cause the display to change state, becoming darker on the viewer side and lighter on the back (sensor) side. As a result, the resistance of the photo-sensor, PR1 , will slowly fall, reducing the voltage on Node 3. As the sensed side approaches the desired value, the value of PR1 rises until equilibrium is reached at the desired grey level. As before, this system can be operated with the sensing illumination present for all of the time or, with the addition of a further TFT equivalent to T5 in Figure 5, in a pulsed mode where the illumination is only present for part of the time. There are a range of variations possible on the detailed design and drive of such a circuit, but these examples of circuit have the defining feature of a reset phase followed by a drive phase during which the photo-feedback loop is operative to ensure the correct grey level is achieved.

The circuits described above use an inverter type output stage to drive the electrophoretic display. As outlined above, when the inverter is made with NMOS technology (as is available for a-Si), the inverter stage can draw significant current during operation. One way of alleviating this outlined above is to use a pulsed operating mode, but this requires an extra TFT in the pixel and more complex waveforms. The inverter stage also has a limited gain, so the electrophoretic drive voltage is reduced as the target brightness is approached. A further modification allows the full drive voltage to be maintained on the electrophoretic cell until the desired brightness is achieved. The drive voltage can then be turned off rapidly ('snap-off') to fix the pixel brightness in the desired state.

The pixel circuit of this further modification is illustrated in Figure 10.

This corresponds to the circuit of Figure 2, but the lower voltage rail (VP2) is ground, and is common with one terminal of the display pixel. In this case, the transistor T3" functions as a precharging device, and the transistor T4" functions as a snap-off device. The transistor T3 changes the optical state of the display pixel element independently of the applied input pixel voltage, and the snap off transistor positively drives the display pixel element to the stable display state at a time dependent on the applied input pixel voltage and an output of the optical feedback.

The plots in Figure 11 correspond to those in Figure 4.

At the end of the addressing period (which is the same as in previous examples), the voltage at Node 1 (the input to T4") is given by:

V1 = Vdaf + V2 = Vdaf + (VP). R(R2)/(R(PR1 ) + R(R2))

Again, V1 and V2 are the voltages at Nodes 1 and 2.

Essentially, V3 (the voltage at Node 3)is charged to the high power rail VP during the address phase, and remains there until Node 1 reaches the approximately the threshold voltage of T4". When this point is reached, T4" starts to conduct, and Node 3 is discharged, thus preventing further brightness changes in the pixel. Thus the drive signal (Node 3 voltage) is 'snapped-off when the desired brightness is achieved. The circuit as described can only drive the pixel reliably in one direction

(that with positive voltage on the pixel for the case of the n-type a-Si circuit illustrated), so prior to the addressing phase, it is necessary to reset all pixel of the display into a known state, which may be black or white depending on the details of the display effect. The circuit can achieve this by changing the power line voltage to a negative value, and pulsing all rows (either in series or all together). In an alternative addressing scheme using the circuit of Figure 10, it is possible to avoid the need for any reset phase.

In a first driving phase, the power line VP can be negative (rather than the usual positive state). Then the circuit can be addressed in the normal way, but with a shifted column voltage for the same grey level target.

By inverting the power line voltage, the effect on the voltage at Node 2 with changing grey level is inverted. For example, if in the first case (VP positive) Node 2 rises in voltage as the pixel becomes brighter, when VP is inverted, Node 2 will now become more positive as the pixel becomes darker (as VP has been inverted, the direction of the change of the pixel is inverted also). Thus, by programming with an appropriate data voltage from the column, a snap-off action can also be achieved when driving in either direction.

When used in this inverted mode (VP negative for an NMOS circuit) the snap-off threshold of T4 is relative to the voltage on Node 3, rather than a well defined power rail, as in the previous case. This is susceptible to change for example through leakage in the display pixel. Also, the column voltage offset in addressing the two phases of the circuit is now approximately equal in magnitude to the power line voltage VP, which may require a large column voltage swing. This circuit may be used if the frame time is kept short enough that leakage effects in the display pixels are small, so that the circuit may need to be addressed many times to achieve proper operation. If leakage occurs in the pixel voltage, the circuit will tend to overshoot the desired grey level in this phase. This does not matter significantly, provided the circuit is then operated in the normal mode of operation (described before), where the accuracy is significantly higher.

Thus, the circuit is operated to change the pixel state in one direction in a first phase and then in the other direction in a second phase, so that pixels can be driven in either direction to their new optical state. If there is overshoot in the first phase, this is corrected in the second phase, but no overshoot takes place for pixels which are drive by the second phase. An additional factor leading to overshoot in this inverse operating mode is that the snap-off point is effectively relative to the voltage on Node 3, so that as Node 3 starts to discharge, the snap-off TFT tends to turn off again..

Thus, to achieve an arbitrary grey-grey transition accurately, first it is possible to address with negative VP, which drives those that need to be darkened to a darker state, but with the possibility of overshoot, and then drives with positive VP, which drives all pixels lighter, with an accurate turn off when the desired level is achieved.

The description above concerns a vertical electrophoretic display effect, responsive to transverse electric fields. The circuits can also be used with a lateral electrophoretic display effect, as is shown in Figure 12.

The top part of Figure 12 shows a reflective pixel and shows the particles 120 and the container area 122. As shown, ambient light reaches the photosensor, and the quantity is dependent on the pixel setting. Thus, the photosensors are arranged to measure the amount of ambient light that passes through the cell.

The bottom part of Figure 12 shows a pixel used in a backlight mode

(for a transmissive display). A slight scattering of the backlight illumination on the front surface of the display will lead to an illumination of the photosensors even if there is no ambient light, and again the amount will be dependent on the pixel setting. The circuit can thus still operate.

Figure 12 shows to the right the pixel in plan view. As shown, the photosensor only occupies a small part of the pixel aperture, and there are two photosensor devices so that the arrangement can be independent of the absolute light levels (of ambient light or of the illumination source) as described above.

It will be appreciated therefore that the optical feedback system may use a dedicated illumination source, or this may not be required if ambient light is used. Figure 13 shows a possible implementation of a CMOS version of the circuit of Figure 10, such as could be made in an LTPS process. The key advantage of this is that no reset phase is needed, and the circuit can drive directly from grey level to grey level accurately, for both upwards and downwards transitions. In contrast, the circuit of Figure 10 can only drive in one direction accurately, so generally pixels will need to be driven both ways to achieve an accurate result. The drive circuit comprises a first drive transistor (T3a) and a second drive transistor (T3b) each connected between a respective power line and the display pixel element, and the reset portion comprises a first reset transistor (T4a) and a second reset transistor (T4b) in parallel with the display pixel element. The first and second reset transistors (T4a, T4b) are of opposite type.

The circuit has a top portion for changing the pixel state in one direction and a bottom portion for changing the pixel state in the other direction, but this time both have accurate snap-off characteristics. Each portion is addressed independently, and this requires two row control lines, "Row a" and "Row b". One disadvantage of circuit in Figure 13 is the need for a CMOS process. There are many possible variants to the circuit in Figure 13, but most significantly, there is one n-type snap-off TFT (T4a), and one p-type snap-off TFT (T4b). The TFTs T1a-T3a and T1 b-T3b could be of either type, given suitable adjustment of the drive wave forms, and each group (T1a-T3a and T1 b-T3b) could even be of different types. This last option could be used to allow a single row line, if the n-type and p-type threshold voltages were high enough.

The upper part of the circuit of Figure 13 can be operated much like the circuit of Figure 10. In addition, the lower part can be operated in a complementary fashion, to drive transitions in the opposite direction. If a given pixel is addressed by Row a, but already is at or exceeds the target grey level, the circuit will snap-off rapidly, giving little effect. When the pixel is later addressed by Row b, if it exceeds the target grey level, it will be driven downwards until the target is reached, and then the lower circuit is snapped off. The column voltage for the same target grey level may be different for the operation by the two parts of the circuit, with an offset roughly equal to the difference between the n-type threshold voltage and the p-type threshold voltage.

Thus, a display may be updated by addressing all pixels via rows a, and waiting such a time as all have snapped off. Then all pixels are addressed by rows b, and the same occurs. An alternative would be to update via each set of rows alternately, in frames that are long compared to the snap-off time, but short compared to the rate of change of the display state. This would have a more pleasing visual effect, as all transitions would appear to happen simultaneously, but require more updates. Figure 14 shows a modification to Figure 10 to utilise a faster snap-off of the voltage across Node 3. The reset (snap-off) portion comprises a plurality of cascaded transistors T6, T4", one of which is in parallel with the display pixel element. In the example shown, a first reset transistor T6 charges the gate of the output transistor T4" which is in parallel with the display pixel element.

In the example of Figure 10, the voltage at Node 1 rises for a positive transition on the electrophoretic pixel, and the snap-off TFT T4" takes a finite time to switch on as the desired brightness is reached. This can lead to errors in brightness. In Figure 14, the snap-off TFT remains off held by the voltage stored on its gate, either by an additional capacitor or the gate-source capacitance (if this is sufficient). Then, as the threshold of T5 is reached, the voltage on Node 4 will start to rise, with a current gain of G_m, allowing T4" to switch on much faster and limit the error voltages introduced by threshold variations. This circuit will provide improved brightness uniformity and speed up the snap-off time.

This circuit can be adapted to have a bi-directional mode of operation, in the same way as the circuit of Figure 10. One way to do this is that the power supply to T5 and T6 can be attached to separate switchable lines, so they can be changed independently of the VP voltage. However, the circuit will be inaccurate when driving in the negative direction, for the same reasons as described above. For this reason, it could again be driven with a negative phase, followed by a positive phase to achieve an accurate result. The examples above relate to an electrophoretic display device. The invention can be used in other bistable display devices using active matrix addressing, for example electrochromic displays.

Electrophoretic display systems can form the basis of a variety of applications where information may be displayed, for example in the form of information signs, public transport signs, advertising posters, pricing labels, billboards etc. In addition, they may be used where a changing non-information surface is required, such as wallpaper with a changing pattern or colour, especially if the surface requires a paper like appearance. The physical design of the pixels has not been described in detail, as this will be known to those skilled in the art.

Various modifications will be apparent to those skilled in the art.

Claims

1. A display device, comprising an array of rows and columns of pixels disposed over a common substrate, wherein each pixel comprises: a display pixel element having a stable drive state in which the optical characteristics remain substantially constant; a drive circuit for controlling the voltage applied to the display pixel element; an optical sensor arrangement (PR1 , R2) for detecting an optical signal dependent on the setting of the display pixel element; wherein the drive circuit comprises an optical feedback control system which is arranged to drive the display pixel element to the stable display state in response to an applied input pixel voltage and an output of the optical sensor arrangement, the optical characteristics being varied before the drive of the display pixel element reaches the stable state and being held substantially constant once the stable state has been reached.

2. A device as claimed in claim 1 , wherein the optical feedback control system implements a function between the optical output of the display pixel element and the applied input pixel voltage.

3. A device as claimed in claim 1 or 2, wherein the optical feedback control system is arranged to drive the display pixel element to a condition in which OV is applied across the display element.

4. A device as claimed in any preceding claim, wherein the drive circuit comprises an input capacitor (C1 ) for charging to the input pixel voltage.

5. A device as claimed in claim 4, wherein the input capacitor is connected at one terminal (Node 1) to a data input line though an address transistor (T1 ).

6. A device as claimed in claim 5, wherein the optical sensor arrangement comprises a light-sensitive potentiometer, and wherein the input capacitor (C1 ) is connected at its other terminal to the potentiometer output.

7. A device as claimed in any preceding claim, wherein the drive circuit comprises an inverter (T3.T4) having as input (Node 1) a signal which is dependent on the output of the optical sensor arrangement and the applied input pixel voltage.

8. A device as claimed in claim 7, wherein an increase in pixel output brightness results in a signal at the output of the inverter which drives the display pixel element to a darker pixel output, and a decrease in pixel output brightness results in a signal at the output of the inverter which drives the display pixel element to a brighter pixel output.

9. A device as claimed in claim 7 or 8, wherein the drive circuit comprises an input capacitor (C1) for charging to the input pixel voltage, connected at one terminal (Node 1 ) to a data input line though an address transistor (T1 ) and wherein the one input terminal (Node 1 ) of the input capacitor (C1 ) is provided to the inverter input.

10. A device as claimed in any one of claims 1 to 6, wherein the drive circuit comprises a reset portion (T4') for driving the display pixel element to a reset state, and a drive transistor (T3') for changing the optical state until the stable display state is reached.

11. A device as claimed in claim 10, wherein the drive circuit comprises an input capacitor (C1) for charging to the input pixel voltage, connected at one terminal (Node 1) to a data input line though an address transistor (T1 ) and wherein the one input terminal (Node 1 ) of the input capacitor (C1 ) is provided to the gate of the drive transistor, the drive transistor (T3') being connected between a power line and the display pixel element.

12. A device as claimed in any preceding claim, wherein an output transistor (T5) is provided between the output (Node 3) of the optical feedback control system and the display pixel element.

13. A device as claimed in claim 12, wherein the output transistor

(T5) is pulsed.

14. A device as claimed in any one of claims 1 to 6, wherein the drive circuit comprises a drive transistor (T3") for changing the optical state of the display pixel element independently of the applied input pixel voltage, and a reset portion (T4") for driving the display pixel element to the stable display state at a time dependent on the applied input pixel voltage and an output of the optical sensor.

15. A device as claimed in claim 14, wherein the drive transistor

(T3") is connected between a power line and the display pixel element, and the reset portion comprises a transistor (T4") in parallel with the display pixel element.

16. A device as claimed in claim 15, wherein the drive circuit comprises a first drive transistor (T3a) and a second drive transistor (T3b) each connected between a respective power line and the display pixel element, and wherein the reset portion comprises a first reset transistor (T4a) and a second reset transistor (T4b) in parallel with the display pixel element.

17. A device as claimed in claim 16, wherein the first and second reset transistors (T4a, T4b) are of opposite type.

18. A device as claimed in any one of claims 14 to 17, wherein the reset portion comprises a plurality of cascaded transistors (T6, T4"), one of which is in parallel with the display pixel element.

19. A device as claimed in claim 18, wherein the reset portion comprises a first reset transistor (T6) for charging the gate of an output reset transistor (T4") which is in parallel with the display pixel element.

20. A device as claimed in any preceding claim, further comprising an illumination source for providing a reference illumination to the optical sensor arrangement.

21. A device as claimed in claim 20, wherein the optical sensor arrangement comprises first and second optical sensing elements, one of which is responsive to the illumination source only and the other of which is responsive to a display pixel setting.

22. A device as claimed in any preceding claim, wherein the display pixel is a reflective display pixel.

23. A device as claimed in claim 22, wherein the optical sensor arrangement detects a light intensity of light reflected from the opposite side of the display pixel to the output side of the display pixel, the reflected light being reflected light of an illumination source.

24. A device as claimed in any one of claims 1 to 21 , wherein the display pixel is a transmissive, reflective or transflective display pixel.

25. A device as claimed in claim 24, wherein the optical sensor arrangement detects a light intensity of ambient light transmitted through the display pixel.

26. A device as claimed in any preceding claim, comprising an electrophoretic active matrix display device.

27. A method of driving a display device, the display device comprising an array of rows and columns of pixels disposed over a common substrate, wherein each pixel comprises a display pixel element having a stable drive state in which the optical characteristics remain substantially constant, wherein the method comprises, for each display pixel: varying the optical characteristics by applying a voltage to the display pixel element; using an optical feedback signal dependent on the setting of the display pixel element and dependent on an applied input pixel voltage to drive the display pixel element to the stable display state, and using in-pixel optical feedback and drive circuitry; and holding the optical characteristics substantially constant once the stable state has been reached.

28. A method as claimed in claim 27, wherein the drive voltage of the stable drive state is OV is across the display element.

29. A method as claimed in claim 27 or 28, comprising charging an input capacitor (C1 ) to the input pixel voltage.

30. A method as claimed in claim 27, 28 or 29, comprising resetting the display pixel before changing the voltage applied to the display pixel element towards the drive voltage of the stable state.

31. A method as claimed in any one of claims 27 to 30, further comprising pulsing an output transistor (T5) in series with the display pixel element.

32. A method as claimed in any one of claims 27 to 30, comprising initially driving the display pixel to a predetermined voltage, and subsequently driving the display pixel element to the stable display state at a time determined by the feedback control.

33. A method as claimed in claim 32, wherein driving the display pixel element to the stable display state comprises actuating a reset transistor (T4') in parallel with the display pixel element.

34. A method as claimed in any one of claims 27 to 33, comprising, in a first phase, varying the optical characteristics in a first direction for those pixels which are to be driven to an optical state which differs from the present optical state in that first direction, and in a second phase, varying the optical characteristics in an opposite second direction for those pixels which are to be driven to an optical state which differs from the present optical state in that second direction.