Advantages and Limitations of the CRT
As might be expected of a display device of such relative complexity, the CRT suffers from a number of limitations and shortcomings. However, it also represents a very mature technology, and one in which these problems have for the most part been addressed. This maturity also results in one of the CRT’s major advantages over other technologies – it is by far the least costly alternative in a wide range of potential applications.
Another major advantage of the CRT display, from a functional perspective, is its ability to adapt quickly and easily to a wide range of image formats. Unlike most other display technologies, the CRT does not have an inherent, fixed format of its own. The number of scan lines in the raster is determined solely by the scanning frequencies used in the beam deflection system. Even in the case of the color CRT, the phosphor triads of the screen do not represent fixed pixels; there is neither a requirement, nor any mechanism provided, to ensure that the samples in the image in any way align with these. This has permitted the development of “multifrequency” CRT displays, capable of accepting a wide range of image formats. Other advantages of the CRT include a wide viewing angle, a reasonably wide color gamut, and the fact that it is an emissive display and so does not rely on any external light source.
Limitations or disadvantages of the CRT include the difficulty in obtaining an acceptable level of geometric distortion, linearity, and focus, plus the aforementioned color problems of misconvergence and color purity. Most of these are complicated by the fact that the distance from the electron gun (or more precisely, from an imagined “point of deflection” at which the beam is thought of as being “bent” onto the correct trajectory) to the screen varies considerably from the center of the screen to the corners. This means that there will be no one fixed focus or convergence correction which will be optimum for the entire screen, and also results in an inherent geometric distortion if a straightforward linear-ramp waveform were to be used for the deflection currents. All of these require compensation in the form of complex waveforms which correct the focus, beam position, etc., vs. the beam’s position on the screen. In modern monitor designs, however, these are relatively easy to produce via digital synthesis.
More difficult to correct, however, are the CRT’s inherent size, weight, and relatively high power requirements. Still, given its low cost and generally good performance across the board, the CRT remains the display of choice in a wide range of applications; the only display needs which have not somehow been addressed by this technology are limited to a relatively narrow range of portable systems, or a few specific environments (such as those with high ambient magnetic fields). CRTs have been built small enough (approx. 2.5 cm diagonal) to permit their use in such applications as viewfinders or head-mounted displays; they have also been used as direct-view displays, as in a monitor or television, up to 1 m in diagonal measurement and even higher. Special CRT types are also used as the basis of projection systems (to be discussed in more detail later in this topic) providing images several meters or more across, and even more specialized relatives of the CRT are used as the basic element making up extremely large display systems for sports arenas and similar venues. After almost 100 years as a practical display device, the CRT is still in widespread use and appears likely to remain a significant portion of the overall display market for many years to come.
The “Flat Panel” Display Technologies
The vast majority of commercially viable, non-CRT display technologies may be referred to generically as “flat-panel” displays, or FPDs. These types have several characteristics in common. First, they are, as the name implies, physically flat and relatively thin devices, certainly with much less overall depth than the CRT. Second, they are fixed-format displays; they are in general composites, a matrix of simple display “cells” or “pixels”, each producing or controlling light uniformly over a small area. Each of the individual cells of such displays is driven such that they correspond one-to-one with the samples or pixels of the image. This is a fundamental difference with the operation of the CRT, and accounts for much of the differences in both the appearance and the interface requirements for the two classes of display.
Since these displays do consist of a fixed array of individual elements, it is natural that they be organized and driven in a manner analogous to the array of pixels resulting from the sampling of the image.In short, these elements are typically arranged and driven in a regular rectangular array of rows and columns. Each element may be accessed, or addressed, simply by selecting the appropriate row and column. This matrix-addressing scheme is common to practically all flat-panel display types. One result is that, unlike the CRT, which readily accepts a continuous video signal, the flat-panel types require that the video information be provided in discrete samples, corresponding to the discrete pixel structure. (This discrete structure has led some to refer to these types of displays as “inherently digital”; this belief is, however, in error. The fixed-format structure places additional demands on the timing or sampling of the incoming information, but does not necessarily require either digital or analog encoding of this information.
A simplified display employing a matrix of row and column electrodes is shown in Figure 4-6. In this case, we are using separate light-emitting diodes (LEDs) located at the intersection of the rows and columns; a given LED will light only when the proper row and column electrode pair is selected and driven. Thus, each LED may be considered as forming one pixel of this display. In most practical displays, however, the active display element will commonly be located between the electrodes, with the electrodes themselves carried on substrates defining the “front” and “back” (or “top” and “bottom”) of the display device. As these are displays, which will either produce or transmit light at each pixel location, at least one set of electrodes must typically be constructed on a transparent substrate, usually glass. In many cases, the electrodes themselves are transparent, made from very thin but still sufficiently conductive layers of metal film or metallic oxides. (One of the most common materials for this purpose is indium-tin oxide, generally referred to as “ITO”.) Drive circuitry is located around the periphery of the panel, such that the electrodes may be selectively energized to drive the desired pixel. In operation, with the video information provided in the previously described raster-scan order, each pixel would be written to the column electrodes while the row electrode corresponding to that line is selected. However, this shows a potential problem with this simplistic drive scheme; each pixel will be active only during that brief period of time in which it is being driven, and will return to its inactive state as soon as the next pixel is selected. In a display with a large number of pixels, it becomes very difficult to drive each pixel long enough to ensure that it is at least seen as in a stable state.
Figure 4-6 Row-column matrix addressing. In this simple display, the picture elements (LEDs) are located at the intersection of row (white) and column (dark) electrodes. Driving a given pixel requires simply connecting a source to the row and column electrodes that intersect at the desired location, as shown.
One way to correct for this is to load the video information into a storage element, such as a shift register, and then write all of the pixels in a given row (or at least part of a row) at the same time. This permits each pixel to be driven for a line time, rather than a pixel time, before the next row must be written. With each pixel enabled for a longer time, the display can appear more stable (with higher brightness and contrast) to the viewer. There is still a limit to how far such a drive scheme can be extended, however; with increasing line counts, the “on” time for each pixel may again be decreased to the point at which the display would not be usable. Increasing the pixel count and/or the physical size of the display also leads to increased capacitance in the row and column electrodes, making it more difficult to drive them quickly. This is especially a problem with those display technologies which require relatively high voltages and/or currents.
The basic problem faced here is the same as in the CRT; any given area needs to be driven long enough, and/or with sufficient intensity, to register visually, and must repeatedly be driven or refreshed so as to create the illusion of a steady image. But increasing the pixel count (or the physical size of the screen) also increases the difficulty of achieving this. In the case of the CRT, the problem is partially ameliorated through the persistence of the phosphor; it continues to emit light for some period of time after the direct excitation of the beam is removed. Some of the flat-panel technologies provide similar characteristics (in some cases, exactly the same: phosphor persistence), but a more common solution is to design “memory” into each pixel. In other words, the panel will be designed such that each pixel, once addressed and driven, will maintain the proper drive level on its own. This is typically achieved by constructing a storage element, usually comprising at least a transistor and a storage capacitor, at each pixel location. Flat-panel displays employing such schemes are generically known as active-matrix displays, due to the active electronic elements within the pixel array itself, while the simpler system in which the electrodes drive the picture elements directly become the passive-matrix displays.
The fact that these displays are of a fixed pixel format is, again, one of the chief functional differences between this class and the highly flexible CRT. There are a certain fixed number of pixels in an FPD, and so it must always be driven at its “native” format. In order to use a flat-panel display in applications that traditionally have used the CRT – such as computer monitors – it is often necessary to add intermediary circuitry which will convert various incoming image formats to the single format required by the display. Such image-scaling may be done using a variety of techniques, some more successful than others in providing a “natural-looking” image. The FPD also, again unlike the CRT, may be restricted to a relatively narrow range of frame rates, requiring also that frame-rate conversion be provided for even if the input is of the correct spatial format. This can again result in differences in image appearance between the CRT and FPD displays, especially if moving images are to be shown. Finally, even if no spatial or temporal conversions of the input image are required for display on the FPD, the simple fact that its pixels are of a fixed and well-defined shape results in a significant difference in appearance between the image on an FPD-based monitor and the same image as seen on a CRT.
At this point, we review the fundamental operation of several of the more popular non-CRT technologies. While this is certainly not be an exhaustive, detailed description of all FPD operating modes, it should serve to give some idea of the wide and varied range of types which are offered under this general classification.
By far the most common of the flat-panel display technologies is the liquid crystal display, or LCD. Now used in everything from simple calculator, watch, and control panel displays to sophisticated full-color desktop monitors, the LCD is almost synonymous with “flat-panel display” in many market at present.
Unlike the other flat-panel types to be reviewed here, the LCD is a non-emissive display. It acts only to modulate or switch an external light source, either as that light passes through the LCD (the transmissive mode of operation), or as the light is reflected from the LCD structure (operating in reflective mode). There are numerous specific means through which LCDs control light, but all operate in the same fundamental manner – the arrangement of molecules within a fluid is altered through the application of an electric field across the material. The effect on the light transmission or reflection may be through phase or polarization changes, the selective absorption of light, or by switching between scattering and nonscattering states.
Probably the most common operating mode, and certainly one of the most useful in explaining the basic of LC operation, is the twisted-nematic mode. Liquid crystals are so named because the molecules of the liquid tend to align themselves in ordered arrays, as in a solid crystalline substance. These materials are also generally organic compounds in which the molecules are relatively long and thin; for the purposes of analyzing their electro-optical behavior, they may be though of as extremely small rods in suspension in a fluid medium. In the nematic state, these molecules – the “rods” – align themselves in layers throughout the fluid, and such that those in adjacent layers tend to be oriented in the same direction. The molecules will also align themselves with fine physical structures in the substrate of the display. (In practice, these are created by physically rubbing a relatively soft layer of material deposited on top of the glass substrate, creating a very large number of very fine scratches, all aligned in the same direction.) If the liquid crystal material is placed between two such substrates, the tendency of the molecules to align themselves with those above and below, plus the tendency of the outermost layers to align with the “rubbing direction” of the substrate, a sort of helical arrangement of the molecules through the liquid crystal occurs, as shown in Figure 4-7a. This helix has the effect of twisting the polarization of light passing through it by 90°. If crossed polarizing layers are then placed on either side of this structure, light can still pass through by virtue of the polarization rotation.
In Figure 4-7b, however, the effect of placing an electric field across the material is shown. The LC molecules’ tendency to form the helical structure described above can be overcome by a field of sufficient rubbing directions orthogonal to one another, the tendency of the molecules of a given layer to align with those above strength, and the molecules will then instead align themselves with the field. This destroys the helical structure, and with it the polarization rotation effect. Light that previously passed through the second polarizing layer is now blocked. Removal of the electric field permits the helical structure to re-form, and light once again will pass through. The transition between the two states is not especially abrupt, as may be seen in the graph of light transmission vs. applied voltage for a typical LC cell, in Figure 4-8. This gives the TN LCD the inherent capability of producing a range of intensities, or a “gray scale”, although the shape of the response curve is less than ideal.
Figure 4-7 Basic twisted-nematic (TN) liquid-crystal operation. In the off state (a), with no electric field applied across the cell, the liquid-crystal molecules align with each other and with the “rubbing” direction on both substrates. With the substrates crossed as shown, the molecules then form helical structures; this is the twisted-nematic state. This helical structure is also optically active, and will twist the polarization of light by 90° as it passes through the cell. If the substrates also carry crossed polarizing layers, this polarization-rotation action will permit light to pass through the cell, and thus this example is transmissive in the off state. However, if an electric field is applied across the cell, the LC molecules will align with the field, destroying the helical structure and thus eliminating the polarization rotation. Thus, light polarized by the bottom polarizer will not pass the upper, and the cell appears dark. This change of state is completely reversible, simply by removing and applying the electric field, and so will form the basis for a practical display device.
It should be noted at this point that the action described above depends solely on the magnitude of the electric field across the LC cell, not on its polarity; in other words, the liquid crystal display would operate as shown with the source connected in either direction. This turns out to be very important, as it was discovered early in the commercial history of LC displays that the display would be damaged if exposed to a long-term net DC voltage across the cells. This is due to
Many simple liquid-crystal displays are of the passive-matrix type. However, to provide sufficient contrast, the LC materials and cell design used for these result in relatively slow operation. This is necessary so that the individual pixels will remain in the desired state long enough between drive pulses, but makes this type ill-suited to applications requiring the display of rapid motion. Use of an active-matrix design enables faster response, and can result in an LCD suited to motion-imaging applications. Most LCD panels used in high-end applications, such as desktop monitors and notebook computers, are of the active-matrix type, also known as “TFT-LCD” (for “thin film transistor liquid crystal display”; the active com-ponents are constructed via thin films deposited directly onto the display substrate). However, in addition to the added complexity of the active-matrix pixels, this type generally requires more power than the passive-matrix LCDs, making the passive-matrix often the more attractive choice in power-critical portable applications.
Figure 4-8 Idealized response curve of an LC cell. In this case, the cell has been designed to pass more light with the application of an electric field, the opposite of the case shown in Figure 4.7. Both types are in common use.
The simple TN-LCD also suffers from a limited viewing angle, meaning that the appearance of the display is optimum only through a certain limit range of angles, centered around a line roughly perpendicular to display surface. (It should be noted that in almost all practical LC displays, the direction of maximum contrast will not be precisely normal to the plane of the display.) This results from the nature of the electro-optical effect behind the operation of the display, which clearly functions best along the axis of the helical arrangement of molecules. Light passing through the structure at an angle does not experience the distinct change in transmission states, and so the contrast of the display falls off rapidly off-axis. This can be compensated for, to some degree, through the addition of optically active film layers on top of the basic TN panel, or through the use of different LC modes. In the passive-matrix types, the most common approach is to employ the “super-twisted nematic”, or “STN” mode. Without going into unnecessary detail, this mode involves a 270° twist in the helical arrangement of the molecules, rather than the 90° of the standard TN, and provides both higher contrast and a wider viewing angle, along with a much sharper response curve.
Active-matrix LCDs may also use other LC modes rather than the simple TN (with or without compensating film) in order to obtain improved contrast and viewing angle. Two of the more common in current displays are the in-plane switching, or IPS type, and the vertical linear alignment (VLA) mode, both shown in Figure 4-9. These modes are not used in pas-sive-matrix displays, due to their requirement for more complex pixel structures and/or higher power requirements, both of which are contrary to the low-cost/low-power aims of most passive-matrix designs. Both offer greatly improved viewing angle and response times over the conventional TN mode. However, the higher power requirement has limited their use to date to panels intended for desktop monitor or television applications (as opposed to notebook PC applications, which are of course more power-critical). More recently, both types have evolved into “multi-domain” variants; these address color and contrast uniformity issues in the original IPS and VLA types, which resulted from the fact that the LC molecules do not actually swing exactly 90° between states as shown in Figure 4-9. The multi-domain solution is shown in Figure 4-10, using the vertically aligned type as an example. In this approach, the display area is broken into many small areas, each with a different orientation of the LC molecules as shown. When viewed at a normal distance, the color errors introduced by each domain, as viewed from a given angle, cancel each other and the display appears uniform on average.
Figure 4-9 The in-plane switching (IPS) and vertical linear alignment (VLA) LC types. The IPS (a) uses LC molecules aligned in the same direction between coplanar electrodes, rather than the helical arrangement of the TN type. When an electric field is generated between the electrodes, the molecules rotate to align with the field. This 90° (approximately) rotation may also be used to control light passing through the cell, based on polarization. In the VLA type, the molecules are aligned vertically in the off state, but when the field is applied the alignment changes as shown.
Both active- and passive-matrix designs may be used in either transmissive or reflective displays. Transmissive-mode displays most often incorporate an integral “backlight” structure, as shown in Figure 4-11a. The light source itself may be one or more small fluorescent tubes (most often of the cold-cathode fluorescent, or CCFL, type), LEDs, or an electroluminescent panel. To provide acceptable brightness uniformity, some type of diffusing layer is generally also included. The backlight, along with the additional power supply generally required to drive it, again increases the complexity and cost of the complete display system, and so may limit the applicability of such displays to relatively high-end applications. In the reflective LCDs (Figure 4-11b), ambient lighting is used to view the display; rather than a backlight, a reflective layer is placed “behind” the LC panel (as seen by the viewer). Due to the light losses involved in two passes through both polarizing layers and the LC material itself, reflective displays generally provide poor contrast compared to their backlit transmissive counterparts, but still are often the preferred choice where low power con-sumption is of paramount concern. A hybrid type, the transflective display (Figure 4-11c), typically adds a limited-use backlight to a normally reflective display, to enable occasional use in low-ambient-light environments.
Figure 4-10 Multi-domain VLA. To compensate for the non-uniformity of the display if the VLA mode is used, the display may be divided into multiple small domains (at least two per pixel) which differ by having opposite pretilt angles. This is achieved by adding small protrusions to the lower substrate; the opposing domains result in a uniform appearance when viewed together.
Figure 4-11 Transmissive, reflective, and transflective LC displays. In the transmissive type (a), the light source (backlight) is located behind the LC panel itself, and is typically comprised of fluorescent tubes, an electroluminescent panel, or LEDs. In the reflective type (b), common in low-power applications, ambient light is used, passing through the panel from the front and then being reflected back through via a reflective surface behind the panel. The transflective type (c) is a compromise, combining elements of both (a) and (b). Primarily used in the reflective mode, the reflector is made to pass some light from a backlight (usually by making the reflector from a mesh-type material), which is turned on only when insufficient ambient light is available.
Making the LCD into a full-color display is conceptually very simple. With the exception of certain LC modes which involve wavelength-specific effects, this type of display has little or no inherent color, instead passing or reflecting an external light source essentially unchanged. In order to make a full-color display, then, all that is required is the addition of color filters over the LC cells, and the use of a white light source. Various pixel layouts have been used in the design of color panels, but one typical arrangement is simply to place three complete pixel structures – now becoming the three primary-color sub-pixels – into a single square area that is now the complete full-color pixel. Besides the additional complexity in the panel design (which now has at least three times as many “pixels” as in a monochrome panel of the same format), the fabrication and alignment of the color filter layer adds considerable cost to the display.
An alternative method of producing a color LC display is to employ three stacked panels with filter layers corresponding to the subtractive-color primaries (cyan, magenta, and yellow). As a reflective display, this permits full-color operation by selectively absorbing these primary colors.
The term “LCD” covers a wider range of specific technologies than any other of the flat-panel types. There are a very wide range of liquid-crystal types and operating modes which have not been covered in detail here, with varying advantages, disadvantages, and unique features. Some provide very high contrast; some provide bistability, and with it the ability to retain an image even after electrical power is disconnected from the display. However, LCDs have until very recently generally been limited to small-to-medium sized applications; from roughly 2.5 cm (1 inch) (or less) diagonal up to perhaps 63 cm (25 inches) at the upper end. The larger sizes are almost exclusively the domain of the active-matrix types, and the size is for the most part limited by the ability of manufacturers to process sufficiently large panels while maintaining acceptable uniformity and defect counts. There has, however, been some success demonstrated in tiling LCDs, using panels specifically designed to be placed adjacent to one another in order to form a much larger complete display system.