In the electronic process known in the 21st century this is known as analogue techniques and digital techniques where the two techniques are mutually continuous because analogue techniques are techniques with decimal numbers (10 - an = decimal) as well as both trigger and clock processes while digital processes are also formed by a faster process technique such as turning the palm of the hand and removing the result (10 at 2- binary). both are passive and active processes in insert techniques and auto insert in electronic engineering for me when I practice and provide solutions in electronic engineering in my field of work. when we make conjunctions and combine both analog and digital techniques this requires a mathematical calculation process as well as physics and chemistry that are more complicated than usual especially when we want to make electronic components with very super high-tech capabilities. analog and digital when combined in an electronic component is a power that is good enough to support the DIREW process (Data - Input - Read and Write) and REWIND (Read - Write - In - Display) with high quality super electronics . indeed one day in the future we will find a process of electronic engineering that is better than analog and digital processes, but in the 21st century and the beginning of the 22nd century, it is enough to combine and connect Analog and Digital processes. usually at this time there are 7 stages in the process of earning high-quality electronics engineering, 7 stages of the route as follows:1. Electronic inputs or sensors can be electronic components
automatically or switch the power request on the possible input system
Many apply in European continental countries, the United States, Russia
can get extreme cold and extreme temperature extremes.2. Which CPU Processor is connected and combined with
power amplifier and a reset and set system of audio and moving images.3. Output or output transducer can also be called an actuator, the process
electronic engineering of action and reaction styles in various Mathematical theories,
Physics and biology.4. Components of programmed evaluation systems both manual and automatic dynamic motion
and we should form this evaluation system component in panel form
a distributed panel of good and electronic power management systems
also attracts high durability.5. a power supply component that provides energy to all systems
that is to the sensor as input, to the CPU processor, to the Output
actuators and to analogous dynamic motion programs combined digital distributed using discrete mathematics and
linear math (numerical analysis).6. Analog switching components that are analog and digital connected where they are
on all electronic calculation techniques.7. The cable, wireless and dynamic dynamic motion components are in the clock route.thus I described the system of discussion of conjunctions and connections to analog and digital electronics in the time of the 21st century and the beginning of the 22nd century.
Line On Victory Endless = LOVE
Sign in: Gen. Mac Tech for WET ( Work --Energy -- Time )
Electronic displays: a triumph of physics, chemistry and engineering
Electronic displays are an important
feature of modern life. For many years the cathode ray tube, an
engineering marvel, was universal, but its shape was cumbersome and its
operating voltage too high. The need for a flat-panel display, working
at a low voltage, became imperative, and much research has been applied
to this need. Any versatile flat-panel display will exploit an
electro-optical effect, a transparent conductor and an addressing system
to deliver data locally. The first need is to convert an electrical
signal into a visible change. Two methods are available, the first
giving emission of light, the second modulating ambient illumination.
The most useful light-emitting media are semiconductors, historically
exploiting III–V or II–VI compounds, but more recently organic or
polymer semiconductors. Another possible effect uses gas plasma
discharges. The modulating, or subtractive, effects that have been
studied include liquid crystals, electrophoresis, electrowetting and
electrochromism.
A transparent conductor makes it
possible to apply a voltage to an extended area while observing the
results. The design is a compromise, since the free electrons that carry
current also absorb light. The first materials used were metals, but
some semiconductors, when heavily doped, give a better balance, with
high transmission for a low resistance. Delivering data unambiguously to
a million or so picture elements across the display area is no easy
task. The preferred solution is an amorphous silicon thin-film
transistor deposited at each cross-point in an X–Y matrix.
Success
in these endeavours has led to many applications for flat-panel
displays, including television, flexible displays, electronic paper,
electronic books and advertising signs.
Electronic displays are an important
feature of modern life. They have a magical allure, for they make
complex operations within electronic circuits visible. The father of all
displays, the cathode ray tube (CRT), was invented at more or less the
same time as the vacuum valve, and as the valve improved in performance,
and as circuits grew more complex, so visualizing the outputs became
essential, and here the CRT was supreme. The invention of television
(TV) made more demands on the output device, and the CRT developed into
an engineering marvel, giving a multi-colour representation of any
scene, still or moving, that matched the original in fine detail.
Moreover, this device was made in such numbers and at such a low cost
that it could be found in most homes in the developed world.
However,
the design of the CRT gave some serious problems. The first was the
problem of making big screens. It is difficult to bend electrons through
wide angles, so a large-area display had to incorporate a long tube for
the electron source. Further, because thick glass is needed for the
face plate, in order to withstand the air pressure, large tubes are very
heavy, a set with a 36 inch diagonal screen weighing over 100 kg. Other
negative features are the need for high-voltage operation to energize
the phosphors, the risk of X-rays, and the life limit of perhaps 5
years. Scientists were conscious of these limitations and there were
many attempts to invent simpler forms of CRT. Particularly noteworthy
were the tubes proposed by Ross Aiken and by Denes (later Dennis) Gabor,
with the electron gun at the side of the screen. These gave the system a
smaller volume, but could not avoid the other disadvantages.
At
this time, the 1950s and 1960s, electronics was experiencing a
revolution. The invention of the transistor had been followed by the
integrated circuit, with several thousand devices packed into a square
centimetre, and all driven by a few volts. The CRT was becoming more and
more exposed as a dinosaur. The major potential return from the
invention of a display that matched the virtues of the new electronics
induced many companies to undertake research in the field. Defence
laboratories, already anxious to exact real benefits from the simplicity
and reliability of integrated circuits, were also conscious of the need
to find a solid-state display. A simple Doppler radar set could work
off a few volts, and could be held in the hand (figure 1); while the electronic display to show the output needed a high voltage and much more space.
Display requirements
There are three basic requirements for any flat-panel display capable of showing a high-resolution colour picture. They are:
- — an electro-optical effect;
- — a transparent conductor;
- — a nonlinear addressing component.
The most useful
light-emitting media are semiconductors, historically exploiting III–V
or II–VI compounds, but now, more successfully, organic or polymer
semiconductors. There are persistent attempts to breathe new life into
the flat CRT, and much effort into maintaining interest in gas discharge
tubes. Modulating, or subtractive, effects are legion, and the
literature abounds with new candidates, or revisions of old failed ones.
Those that have shown success or potential include liquid crystals,
electrophoresis, electrowetting and electrochromism, in both solids and
liquids. Effects that have not survived include dichroism, whereby
elongated crystals with anisotropic absorption are persuaded to move in
an electric field—crossed crystals absorb, parallel crystals transmit.
The favoured medium was herapathite, quinine iodine sulphate, used in
the first sheet polarizers. This was named after a Bristol doctor,
William Herapath, who observed that when one of his students, a Mr
Phelps, added iodine to the urine of a dog that had been fed quinine,
green crystals grew in the ‘reaction liquid’ (Land 1951).
History does not record the later career of such an imaginative
student, but, alas, it does record the failure of dichroic displays.
The
need for a transparent conductor is obvious, since one needs to apply a
voltage to an extended area while observing the results. The current is
carried by the free electrons, but these also absorb incident
radiation. There is then a balance in the design, because, as the
thickness of the coating is increased to lower the resistance, so the
transmission decreases. The first materials used were metals, and the
top electrode in early selenium photocells was a very thin gold film.
For all metals, the absorption is so high that the electrode can have a
thickness of only a few hundred ångströms, and it is then difficult to
deposit a film of uniform thickness over a large area. Some
semiconductors, when heavily doped, give a better balance, with high
transmission for a low resistance.
Delivering picture
or symbol data to the display is no easy task. Each picture element, or
pixel, must be addressed individually, without crosstalk, and a
large-area TV screen can have well over a million pixels. It would be
impracticable to have two connectors to each pixel, and a matrix of
orthogonal X and Y lines is always used, with the cross-points labelled as an address. The technique is called matrix addressing (figure 2).
Applying half the required voltage to, say, column 2 and row B would
certainly excite solely pixel 2:B, but the simultaneous excitation of
pixel 5:E would also switch on pixels 2:E and 5:B.
Matrix addressing
The remedy is sequential activation, or multiplexing, running down the X rows one by one, applying the half-voltage to the appropriate Y
columns at the correct time. While this does solve the ambiguity
problem, it creates a crosstalk problem, in that a number of
half-voltages will be applied to pixels that should have been passive. Figure 3a
shows that, whenever a pixel is addressed, there is a possibility of a
cross-bar pattern emerging. The situation would be tolerable for a small
display exploiting an electro-optical effect with a voltage threshold,
as in curve A (figure 3b),
where a half-voltage impulse has no effect at all, but most useful
effects have an inadequately sharp threshold, showing a fairly gradual
change in transmission with voltage, as in curve B. The solution is to
add to each pixel a circuit element that does have a sharp threshold,
such as a semiconductor diode or a transistor, converting the passive
matrix into an active matrix (figure 4).
The element must, of course, be integrated on the panel, in the form of
a thin film. Early panels used CdS or CdSe thin-film transistors
(TFTs), but these proved unstable and short-lived. Progress on
flat-panel displays burgeoned only when amorphous silicon TFTs became
available.
Matrix addressing problems. (a) Multiplexing. (b) Electro-optical effects with (curve A) and without (curve B) a threshold.
The active matrix.
Transparent conductors
The
display medium needs two electrodes, and at least one of these must be
transparent, so that the optical path from the pixel to the eye will
suffer little absorption. For most practical effects, the pixel
electrodes are orthogonal to this path, and run the full width of the
display. The conductivity of the electrodes must then be high enough to
give a uniform voltage over the display area.
Let us
assume that we will use a semiconductor for the electrode. Drude theory
shows that free carrier absorption will give an absorption coefficient α at a wavelength λ as
3.1
For a film of thickness t cm with absorption A less than about 20 per cent,
3.2
The resistance per square (Rs) is given by
3.3
3.4
Some 30 years ago it was discovered that cadmium oxide had a low ARs
product, and it was widely used. It was in turn replaced by tin oxide,
which then became a standard. In the 1970s indium oxide doped with tin
(ITO), prepared by spraying, was shown to give better results still, but
today layers are prepared by sputtering, a more practicable technique
for displays. Such layers have been reported with N near 1021 cm−3 and a mobility of 43 cm2 V−1 s−1.
Such a combination is remarkable in a material of such a large energy
gap, particularly when we remember that we are dealing with a
polycrystalline layer. ARs products below 50 have been reported for ITO, but an alternative compound, cadmium stannate, has shown ARs values a factor of 2 lower. Unfortunately, cadmium is highly toxic, and is rarely used in production.
This
does not mean that ITO has no rivals. Though it has excellent
parameters, there are some disadvantages. The high usage of such
coatings has caused the price of indium to escalate, and there is even
the possibility of a world shortage. A more definite disadvantage is the
lack of flexibility in thick coatings, which makes ITO suspect as the
electrode for flexible displays. A possible alternative is ZnO doped
with Al. Sputtered layers have shown values for ARs as low as those for ITO (Igasaki & Saito 1991a,b).
The transparent conducting electrode represents an important achievement for material science that is rarely publicized.
The visual effect
There
is a clear distinction between emissive displays, which work by
generating light, and subtractive displays, which exploit modulation of
incident or transmitted light by controlling absorption, reflection or
scattering. Both classes have their proponents, since each has
shortcomings. Subtractive displays are economic in power, but in the
absence of ambient light a subsidiary source is needed, and much of the
power advantage is then lost. In ambient light, some subtractive effects
can have a dull appearance, lacking charm. Emissive displays are
attractive to view in low or moderate light levels, but they fade and
are difficult to read when bright light falls on the panel. We still
lack an efficient emissive technology, though there has been much
progress in the past few years. To obtain sufficient brightness, it is
necessary to drive the emitting pixels hard, and then their life can be
curtailed.
(a) Subtractive displays
Progress
in semiconductor electronics after 1950 was so rapid that scientists
assumed that semiconductors could do everything. Certainly our
understanding of the physics of solids became deep and wide, and we
could devise strategies for material design to meet a variety of
applications. A number of electro-optical effects in solids were
proposed for displays, but all fell short of the requirements, usually
because they operated at high voltages. It was with some misgivings that
the community realized that the solution could come from liquids, a
previously neglected form of condensed matter, and one where much of our
applied physics and engineering was inappropriate. To augment these
concerns was the discovery that the best effects occurred in organic
materials, previously the almost exclusive province of chemists. Some
relief was felt when it was shown that families of candidate materials
could be produced in weeks, in contrast to the years that were needed to
make and purify inorganic semiconductors like silicon and gallium
arsenide. The years when display concepts were being invented and
refined were truly years of equal partnership between physicists and
chemists.
The 1960s saw four rival subtractive technologies emerging, with little indication that any one had crucial advantages.
(i) Liquid crystals
Over 100 years ago an Austrian botanist, Reinitzer (1888),
discovered a fourth state of matter, with properties intermediate
between solids and classical liquids. This was a liquid that showed
long-range order, though only over a limited temperature range. On
melting from the frozen state, the long-range order made the liquid
cloudy, but as the temperature was increased, it became clear. The
molecules of the liquid were long and thin, and the long directions
followed an ordered pattern. Further research revealed three types of
order. Smectic crystals are similar in their order to crystalline
solids, in that the molecules form equally spaced layers, all pointing
in the same direction, though with little positional order within a
layer. In nematic crystals, there are no layers, the molecules pointing
in the same direction locally, though that direction, called the
director, changes over a distance (figure 5).
Cholesteric crystals are layered, but the molecules lie along the
layers, all those in a layer pointing in the same direction, with that
direction changing gradually and regularly with distance, so that the
director follows a helix.
The director in a nematic liquid crystal.
Liquid crystals (LCs) remained an academic curiosity until 1962, when Williams (1963),
at the Radio Corporation of America’s (RCA) Sarnoff Research Center,
discovered changes in the optical transmission of thin films of para-azoxyanisole
held between two glass slides on the application of 12 V. Williams
subsequently left the laboratory, but his lead was followed by George
Heilmeier, who persuaded the RCA management to start an LC project.
There was, at that time, no room-temperature LC, but the RCA chemists
devised a mixture of three Schiff’s bases that had a nematic range from
22 to 105°C (Goldmacher & Castellano 1967).
The effect that RCA wished to exploit in displays was called the
dynamic scattering mode (DSM), in which the mixture turns from
transparency to white opalescence over a range of a few volts (Heilmeier et al. 1968).
LCs are anisotropic in almost all their physical characteristics. The
values of refractive index, dielectric constant, conductivity,
elasticity and viscosity are very different when measured along the long
molecular axis or the short axes. Because of the dielectric anisotropy,
the molecule will turn in an electric field, and nematics divide into
two classes, positive crystals, which align along the field, and
negative crystals, which try to lie across it. DSM can be generated in
negative nematics, because charges build up along the molecules, giving
rise to a field at right angles to the applied field. At higher fields,
turbulence occurs. RCA publicized their discoveries in 1968, and, amid
some excitement, many companies set about exploiting liquid crystal
displays (LCD) in digital watches and calculators. Curiously, RCA was an
exception.
RCA had little interest in small
instruments. Their display involvement was in CRTs, and here their
position was almost unique. Harold Law and Al Schroeder had invented the
shadow-mask tube some 10 years earlier, and this was now being made in
quantity at plants in Lancaster, Pennsylvania, and Harrison, New Jersey
(Law 1950, 1976). The company had early adopted the principle of licensing their inventions, and shadow-mask tubes were now produced worldwide.1
Actually, this policy was so successful that the royalty income
financed the whole of the Sarnoff Research Center. It was not surprising
that the Development Laboratory did not share the conviction that LCDs
were the way forward for TV displays. They were conscious of the
problems of addressing, and unconvinced that the RCA invention of the
TFT in 1962 (Weimer 1962a,b)
would be the solution. In any case, they did not see the virtue in
trying to replace their own world-leading display, and did not accept
the moral, ‘If you don’t do it, someone else will’.
Looking
back, it is obvious that RCA held most of the assets needed to forge
and preserve world leadership in flat-panel displays, but they opted
out. The management set a target of a 1200 pixel integrated display
panel to be made early in 1968, but when no panel was made by the end of
the year, the project was cancelled. In the next year they abandoned
all work on LC TV, though some work on small displays continued until
1972.
It would not be an overstatement to say that US
industry lost its way on displays in the 1970s. We have seen that the
early running on LCs was made by RCA. That laboratory had not been in
the forefront of discovery on transistors and chips, relying mainly on
licensing from Bell Telephone Laboratories (BTL), but it had a proud
record of innovation in vacuum tubes, culminating in the invention of
the shadow-mask CRT. RCA led in the early research on TFTs and LCDs, but
the belief that flat-panel displays were against their long-term
interests led them to withdraw from the field in 1972. The other
potential industrial leader, BTL, had stayed curiously aloof from the
frenzied search for novel display technology, partly because of their
increased involvement in semiconductor lasers for communications, but
also because senior figures in their laboratory were unconvinced that
new displays were required. They said (Gordon & Anderson 1973):
Prospects for new display technologies are clouded by the fact that there exists a device, the familiar CRT, that has long provided a versatile, elegant, functional, economical, and largely satisfactory solution.
In circumstances where industry was unwilling to lead
in long-term research programmes, defence funding had previously filled
the gap, and we have seen that this indeed happened in the UK. In the
USA, however, the Department of Defense (DoD) was also unconvinced about
flat panels. The opposition was led, strangely, by the scientist who
had contributed much to the earlier enthusiasm at RCA for LCs, George
Heilmeier. He had left RCA in 1970, and within two years was holding a
senior post in the US DoD with responsibility for electronic device
research contracts. He told the main US Displays Conference (Heilmeier 1973):
How many realistic scenarios are there in which we win because we have a flat-panel, matrix-addressed display in the cockpit? We must feed on existing technologies.
It was not surprising then that in the 1970s most of the important developments in this field came from Europe and Japan.
The
lack of management interest in LCDs certainly led to a number of the
RCA scientists leaving, and one of their best theorists, Wolfgang
Helfrich, joined Hoffmann-La Roche (H-LR), the Swiss chemical and
pharmaceutical company, in 1970. There he suggested to Martin Schadt,
the LC group leader, that he should work on a new display effect that
exploited positive nematics. Helfrich’s idea was to make a thin LC cell
that rotated the plane of incident polarized light by 90°. It was known
that nematic molecules would lie down on a glass substrate that had been
rubbed with a polishing cloth in one direction. If that direction was
orthogonal on the two surfaces, a 90° twist would be induced, and when
the cell was put between parallel polarizers, no light could pass.
However, if a field was applied across that cell, the molecules would
align themselves along the field, the twist would disappear, and light
could pass. Schadt made the cell, it worked, and the twisted nematic
(TN) display was born (figure 6).
The twisted nematic display.
There were some curious features to this invention.
Helfrich left RCA in October 1970, made the first TN cells in November,
submitted a patent with Schadt on 4 December (Helfrich & Schadt 1970) and a letter to Applied Physics Letters 3 days later (Schadt & Helfrich 1971).
Such a rapid sequence of conception and construction is unusual. In
fact, as Helfrich admitted 20 years later, he had thought of the TN
effect in 1969, and other ex-RCA staff confirmed this. However, he made
little attempt to attract management interest, since, as he explained,
he was there to help theoretical understanding, not to invent devices.
RCA made no attempt to invalidate the patent or to claim ownership,
possibly because there were further legal complications (Kawamoto 2002).
James
Fergason was a scientist who had worked on cholesteric LCs at
Westinghouse in the early 1960s, but left in 1966 to join Kent State
University. Two years later he formed his own company, ILIXCO, to
manufacture LC displays. In 1968 and 1970 he published two papers that
effectively contained descriptions of the TN display (Arora et al. 1968; Fergason et al. 1970).
He made no attempt then to patent the concept, and was surprised, and
probably irritated, when a colleague reported back after a visit to H-LR
that Schadt and Helfrich had invented a new form of LCD. In fact, it
was as a result of this inadvertent disclosure that H-LR had rapidly
taken the patenting and publishing actions. Fergason himself set about
composing patents and, after an abortive attempt in February, submitted
in April a patent, which was granted in 1973 (Fergason 1971).
No mention was made in this patent of his earlier publications. Though
the validity of Fergason’s patent could have been queried because of
those disclosures, there could be no doubt that he had made and shown a
device in April 1970, because he had recorded the invention in witnessed
notebooks. He therefore had good grounds for contesting the H-LR
patent, and after protracted legal proceedings this was withdrawn.
However, H-LR regained legal ownership of TN rights by buying the
Fergason patent from ILIXCO, which were in financial difficulties. A
compromise agreement shared royalties amicably between all the
interested parties except RCA.
Though the way was now
legally clear for companies to exploit TN displays, the commercial
position was unclear. A number of companies had taken licences from RCA
to exploit dynamic scattering, and they were reluctant to adopt an
untested technology. However, problems soon arose because of the LC
material. DSM effects need negative nematics, and though RCA had now
demonstrated a suitable Schiff’s base that was nematic at room
temperature, it did not have an adequate working range. Sharp developed a
eutectic mixture of three Schiff’s bases that worked over the range
0–40°C, but were then frustrated when their devices failed after only a
few weeks of operation. It became apparent that there was no stable LC
available, and LCDs were acquiring a poor reputation for reliability.
Up
to then, the UK had played little part in LC development, though one or
two university chemistry departments were involved in research, and one
company, Marconi, had patented an LCD before the war (Levin & Levin 1934).
Now events took a curious turn, because a politician became involved.
Much UK semiconductor research had been carried out in government
defence laboratories, and early development of LEDs and diode lasers had
taken place at the Services Electronics Research Laboratory (SERL),
Baldock, and at the Royal Radar Establishment (RRE), Malvern. One of the
aims of the Labour Government elected in March 1966 had been to forge a
‘white-hot technological revolution’, and the next year they
established a Ministry of Technology.
Though formal approval of this recommendation would
normally have taken some months, and, indeed, was never granted, RRE had
anticipated approval, and justified their action on the urgent need for
displays for the portable radar sets they had invented. They
established two consortia, one for materials, involving RRE, Hull
University and BDH, and one for devices, involving RRE, SERL, Marconi,
Rank and STL. The Plessey Research Laboratory at Caswell were also
involved, specializing in electrophoretics. Though most of these
organizations were ‘the usual suspects’ from the semiconductor R&D
programmes, Hull University were unknown. They had come to the attention
of RRE during a meeting held to probe UK expertise on LCs, when it
became clear that Hull, led by Professor George Gray, were clear leaders
in the understanding of LC chemistry. This trust was rewarded manifold.
Gray was given the task of finding a stable LC, because RRE, schooled
in defence requirements for reliable components, appreciated that
consumers also would not tolerate short-lived equipment. All available
LCs had serious shortcomings. Schiff’s bases gave erratic results, and
stilbenes, more recently proposed, were degraded when exposed to
ultraviolet radiation.
The solution did not come
immediately. Hull worked first on carbonate esters, then on sulphonate
and carboxylic esters. They tried purifying samples of Schiff’s bases,
to see if the short device life was linked with impurities, and, when
this failed, moved to stilbene esters and cyano-Schiff’s bases. All
efforts were leading nowhere, and Gray was now becoming frustrated. He
decided to take a step back and see if the materials had a common
fragile feature. Table 1 shows the position in mid-1972.
Table 1.
Early liquid crystal families.
| family | central link | problem |
|---|---|---|
| Schiff’s base | –CN=N– | hydrolytic cleavage |
| stilbene |  | UV instability |
| azo | –N=N– | oxidation, isomerization |
| ester | –CO.O– | nucleophilic attack |
| tolane | –C≡C– | UV instability |
| azoxy | –N=N(O)– | yellow colouring |
Gray
realized that one common feature was the central linking group. It
would be possible to have a biphenyl structure, but this was unlikely to
give a practical working temperature range. Nevertheless, an
appropriate end group might give a reasonable range, and they knew from
their earlier research that –CN gave a strong positive dielectric
anisotropy, crucial for TN devices. The proposed structure was
R
was chosen to be an alkyl or an alkoxy, essentially the cyano-Schiff’s
bases they had worked on earlier, minus the central linkage. After
deciding on the way forward, and making some of the necessary
precursors, in August 1972 Gray and his colleague John Nash left to
attend the International Liquid Crystal Conference at Kent State
University, in the USA.
They left with some reluctance,
for their recently qualified PhD graduate, Ken Harrison, was ready to
attempt the preparation of pentyl-cyano-biphenyl (5CB) and
pentyloxy-cyano-biphenyl (5OCB). They returned to a scene of great
excitement, for both materials had been made and found to be LCs. 5CB
showed a nematic phase from 22 to 35°C, and 5OCB worked from 48 to 69°C.
Even more exciting were the results of stability tests at Malvern. The
resistivity and the transition temperatures of both materials were
unaffected by long exposure to a damp atmosphere, whereas Schiff’s bases
became unusable after a few hours. However, this was just the start,
because now they must find a mixture that met the temperature
requirements, −10 to 60°C. Six alkyls and six alkoxys were then
synthesized, and a number of mixtures of these were tried, but the best
combination had a range only from −3 to 52°C. They needed to design
complicated eutectic systems, but it would have taken far too long to
plot eutectic diagrams for all promising combinations.
A
crucial contribution was then made by Peter Raynes, who had joined the
RRE Displays Group a year earlier, fresh from a PhD on
superconductivity. He realized that the Schroeder–Van Laar equation for
binary eutectics might be extended to predict mixture properties from
thermodynamic data for the individual materials. However, the accuracy
was not high enough, and Raynes then developed a more accurate
semi-empirical method, which proved ideal. This was so useful
commercially that it was not put into print for some years (Raynes 1980).
Melting points of eutectic mixtures were then predictable to within
5°C, and clearing points, the change from nematic to isotropic
behaviour, to within 2°C (Hulme et al. 1974).
Raynes predicted that no mixture of biphenyls would operate well below
zero. Gray then reasoned that adding a terphenyl component would give a
wider range mixture, and though terphenyls were difficult to make, they
proved to be the solution.
Meanwhile, production
processes of pure biphenyls had been perfected at Poole, where Ben
Sturgeon, the Technical Director of BDH, had made inspired
contributions, and before long BDH was selling biphenyl eutectics
widely, for though their temperature range was not ideal, their
stability was very attractive. Late in August 1973, Raynes made a
four-component eutectic that had a range of −9 to 59°C. It was called
E7, and the composition is shown in figure 8. In almost all respects it met the specifications put to RRE by manufacturers of watch displays (table 2).
Composition of eutectic LC material E7.
Table 2.
E7 as a display material.
| property | requirement | E7 |
|---|---|---|
| nematic range | −10°C to approx. 60°C | −9°C to 59°C |
| threshold voltage | less than 1.0 V | 1.3 V |
| resistivity | more than 5×1010 Ω cm | 5×1010 Ω cm |
| response time | less than 50 ms | 35 ms |
| colour | none | none |
only
BDH and H-LR could sell biphenyls. Rapidly they dominated the market.
By 1977 BDH were the largest manufacturers of LCs in the world (figure 9), and biphenyls had become their largest-selling product. Less than five years earlier, the company had never made an LC.3
The market for liquid crystals, 1975–1979.
I
should not give the impression that the biphenyls had no rivals. The
German company Merck Chemicals had made LCs for academic users since
1904, had increased production to meet commercial demand in 1968, and
commanded the market before biphenyls appeared. They did not remain
idle. In 1973 they bought BDH, and, not wishing to disturb the close
relationship between BDH, Hull and the MoD, ran parallel centres in
Poole and Darmstadt. Darmstadt was stimulated by the competition, and
they conceived their own stable family, similar to
alkyl-cyano-biphenyls, but with a cyclohexane ring substituted for one
phenyl ring. This family, known as the phenyl-cyclohexanes (PCH), became
second in importance to the biphenyls, having a lower viscosity and
some very favourable optical properties (Eidenschink et al. 1977).
There
are many physical parameters of LCs that control the electro-optical
behaviour, but the most important for displays are the elastic constants
and the rotational viscosity. Table 3 gives the room-temperature values for E7 for the splay (k11), twist (k22) and bend (k33) elastic constants and the viscosity (η).
Table 3.
Parameters for LC material E7.
| k11 (×10−12 N) | k22 (×10−12 N) | k33 (×10−12 N) | η (cP) |
|---|---|---|---|
| 10.70 | 10 | 20.70 | 38 |
The threshold voltage for switching, VT, is given by
4.1
4.2
It can be seen that VT is independent of the viscosity and the cell thickness (d), but the time constants depend on both. We have
4.3
4.4
4.5
4.6
, a rapid switch-on is readily available at higher voltages.
The
visual appearance of a TN cell depends strongly on the angle of view,
and both the viewing angle and the contrast ratio came under criticism
as the possibility of major markets became apparent. Major advances were
made, both in the cell configuration and in the LC materials. A big
step forward was the idea of increasing the twist from 90° to 270°. This
supertwist nematic display (STN) was proposed and patented in 1982 by Waters & Raynes (1982) at RRE, and independently patented a year later by the Brown Boveri group, led by Terry Scheffer (Amstutz et al. 1983),
afterwards ably assisted by Jurgen Nehring. STN gave the steep
threshold necessary for passive matrix displays, and the response time
and angle of view were similar to the simple TN device (Scheffer 1983; Waters et al. 1983).
It became the preferred display for instruments and lap-top computers,
and lost ground only when the production of TFTs over large areas was
perfected. The STN display was patented and widely licensed by the MoD,
and yielded royalties of over £100 million, the largest return for any
MoD patent.
More radical changes to the TN mode were also introduced. Soref (1972, 1973),
at Beckman Instruments and Sperry Rand, had proposed in 1972 displays
using circular polarizers with interdigitated electrodes on just one of
the glass surfaces. The concept of interdigitated electrodes was
improved by the Freiburg Fraunhofer Institute, which invented the
in-plane switching (IPS) display in 1990 (Baur et al. 1990; Kiefer et al. 1992).
The
electrodes are on the same cell surface, and, in the absence of a
voltage, the LC molecules lie parallel to the surfaces, which have the
same rubbing direction, so there is no effect on polarized light.
Application of a field between the electrodes induces a rotation on that
cell surface, and a twist between the two surfaces. However, fringe
fields and the effect of tilt make the operation more complicated, and
can lead to increased response time. Moreover, each pixel needs two
switching TFTs, and in early versions this reduced the transmittance.
IPS was studied by a number of laboratories in the 1990s, notably
Hosiden, NEC and, particularly, Hitachi (Ohe & Kondo 1995; Ohta et al. 1996). There are now a number of variants in commercial production.
Though
TN mode devices showed clear advantages over dynamic scattering,
several laboratories pursued other LC effects in the 1970s. Fred Kahn at
BTL proposed in 1971 a switching effect based on negative nematics
aligned homeotropically, i.e. at 90° to the cell walls, so that the
electric field was parallel to the long axis of the molecules, the cell
being placed between crossed or parallel polarizers. Application of the
field will then cause the molecules to rotate through 90°, and the
transmission through the cell will change (Kahn 1971, 1972). Kahn showed that VT was given by equation (4.1), with k=k33. For the LCs he used, VT
was 3 V, and the birefringence increased steadily as the voltage was
increased to 20 V. Though this seems a simple mode, the alignment
requirements are severe. The homogeneous alignment used in TN cells is
readily obtained by rubbing the glass surface in one direction. This
creates microgrooves, and the long molecules lie in them. For Kahn’s
vertical alignment (VA) mode, it is necessary not only to persuade the
molecules to lie at 90° to the surface, but also to impose a slight
tilt, to give a source of defined anisotropy. This proved difficult to
achieve over areas practical for displays, and exploitation of VA
awaited the sophisticated control over LC that developed during the next
20 years. A number of improvements were then proposed, one of the most
effective being the Fujitsu multi-domain vertical alignment (MVA) mode (figure 10).
Operating principle of Fujitsu MVA-LCD (after Koike & Okamoto 1999). (a) Basic concept. (b) Basic cell structure.
In
the off state, the molecules adopt a more-or-less homeotropic
structure. When voltage is applied, within each domain the molecules
align orthogonally to the raised structures, which are actually
pyramids, so the light is channelled into a cone, giving a wide angle of
view.
The early thirst for LCs shown in figure 9 has not diminished. In 1979 world sales were £5 million. In 2006 Merck, the dominant supplier, had sales of £500 million (figure 11)
and their provisional figures for 2007 exceeded £700 million. Two other
large suppliers, Chisso and Dai Nippon Ink, are anticipated to have
sales of over £300 million, making a sales total of at least £1 billion
in 2006. In 1979 sales were measured in grams. Today the measurement
unit is tonnes.
Merck sales of liquid crystals.
Naturally,
this growth has had to be served by much R&D on materials to give
better display performance. As I noted earlier, the most important
parameters to consider in designing an LC for displays are the
temperature range of the nematic phase, the magnitude of the elastic
constants, the ratio of the bend constant k33 to the splay constant k11, and the viscosity η. Also relevant are the magnitude and sign of the dielectric anisotropy Δε, and the magnitude of the birefringence Δn.
Though
biphenyls and phenyl-cyclohexanes served the LCD industry well during
the first 15 years of development, there were obvious deficiencies in
the display appearance and multiplexing capability. One serious problem
was the resistivity, insufficiently high for large displays. LCs are
insulating, but that is a relative term, and to ensure that the pixel
voltage does not drain away in active matrix applications, the
resistivity must be very high, above 1012 Ω cm, and that
rules out some otherwise promising families. Another problem was the
slow switching speed, with a failure to follow fast-changing images. The
simple remedy of reducing viscosity led to both a smaller operating
temperature range and a reduction in the dielectric anisotropy, giving a
higher switching voltage. After much research at Hull University and
Merck, the inclusion of fluorine groups was shown to give much improved
performance (Gray et al. 1989; Reiffenrath et al.1989a,b; Coates et al. 1993).
It should be noted that commercial LCs now are mixtures of from 10 to
30 individual molecules, but a typical core material is shown in figure 12. This material has a high Δε of over 17, satisfactory for both IPS and TN modes (Kirsch 2004).
A Merck superfluorinated terphenyl.
The design of materials for vertically aligned nematic (VAN) mode poses new problems, since they must have a large negative Δε.
It was known that lateral polar substituents would lead to a dipole
moment orthogonal to the long axis, and again fluorine was the preferred
substituent. The most useful materials are difluorophenyl derivatives
and more recently the difluorindanes have been invented, giving a Δε as large as −8. VAN switching times are longer than for IPS or TN, and the cell spacing has to be as small as 3 μm.
This, in turn, calls for a larger value of birefringence, and this
often results in high viscosity. A good compromise is found in the
terphenyl shown in figure 13, which has Δε=−2.5, Δn=0.23 and η=90 cP (Pauluth & Tarumi 2004). Additional fluorine substituents give larger negative values of Δε, but the viscosity is increased by a factor of three or more.
(ii) Electrophoresis
The
motion of solid particles in a liquid medium under the action of an
electric field has been studied for many years. It is an easy and
inexpensive technique, and is often used for species separation and for
particle deposition on substrates, as in painting large areas uniformly.
The system becomes a display when the particles have a colour that
contrasts with that of the liquid. Typically, white TiO2 particles in blue ink give a most impressive high-contrast image (figure 14).
Electrophoresis
(EP) was one of the early candidates for flat panels, but research
revealed several major disadvantages. The most serious were
gravitational settling, flocculation of the individual particles into
clumps and electrode staining. Staining appeared in time because a field
sufficiently high to move the particles towards an electrode would not
always be sufficient to extract it. Clumping gave a variation in
mobility, so that a clump would have a slower response time than a
single particle. In any case, the drive voltage was higher than for
LCDs, and the response time slower. Some of the problems were summarized
by Hopper & Novotny (1979) from Xerox.
The
effect was thought unpromising by the display community for several
years until the Nippon Mektron company suggested that microencapsulating
the particles would solve many of these problems (Osamu et al. 1987). There followed refinements of this technique (Jacobson et al. 1997; Comiskey et al. 1998)
until today it is regarded as a leading candidate for electronic paper,
which will be considered later. The more general application to
displays is less obvious, because transmissive effects are poor.
(iii) Electrochromism
Striking
colour changes are seen in some solids and liquids on the application
of an electric field. The first attempt to exploit such effects came in
1948 from Sziklai (1948),
a prolific inventor at RCA, who patented an electrically controlled
colour filter for TV cameras. The realization that this gave the
foundation for flat-panel displays came almost 20 years later, but then a
number of laboratories produced simple prototypes. Early in the field
of electrochromic solids was American Cyanamid (Deb & Shaw 1966; Deb & Witzke 1975) and ICI, which studied effects in organic liquids in-house (Kenworthy 1971) and sponsored research on oxides at Imperial College (Green & Richman 1974; Green 1976). The Philips Research Laboratory, Eindhoven, made practical displays based on viologen derivatives (Van Dam et al. 1972; Schoot et al. 1973).
The
basis of the effect is a reduction–oxidation reaction, and for many
years the favoured materials were either simple salts or polymeric forms
of bipyridyls or, for the ‘solid’ displays, transition metal oxides.
The oxide displays used a fast-ion conductor as the electrolyte. The
most-studied liquid system used viologen derivatives, and operated by
combining a plating reaction with a dye reaction, so depositing a violet
film on the transparent electrode. One advantage of electrochromism is
that it has memory, so that power is required only when the data change.
Unfortunately, this advantage is temporary, since the effect is not
absolutely reversible, and in time the electrodes become stained. Early
viologen displays had a short lifetime because of recrystallization. The
first solid displays used WO3, but the response time was
over a second. Combinations with Mo, V and Nb gave different colours,
and a move to liquid electrolytes improved the response time to some
tens of milliseconds, but the lifetime was too short for general
adoption. There might have been more sustained interest if the effect
had a threshold, but, as it was, multiplexing was difficult, and there
was therefore little advantage over the alternative effects.
New
life has been given recently to the science of electrochromism by the
appearance of two important applications, electronic paper and variable
transmission windows. This interest has stimulated materials research,
and electrochromism in polymers has made much progress. However, the
short life, about 106 cycles, and the long response time of
several hundred milliseconds, remain as problems to be overcome before
electrochromism can play a part in dynamic displays.
(iv) Electrowetting
For
a brief period after 1980 it appeared that the tide on the wisdom of
flat-panel development was turning in BTL as the influence of the
opposing senior managers diminished, and a novel approach was published.
This was based on inducing motion in a fluid by the application of an
electric field. Gerardo Beni and Susan Hackwood showed that a porous
solid could be filled with a liquid of the same refractive index, so
that it was transparent. If the liquid was then extracted from the
solid, the pixel would be diffusely reflecting. They found that visible
changes occurred at applied voltages of about 1 V with response times of
a millisecond (Beni & Hackwood 1981).
This promising start was not pursued for long at BTL, but the
technology was exploited in microfluidics and tunable liquid
microlenses. The circle was completed over 20 years later when Robert
Hayes and Johan Feenstra, of Philips Research in Eindhoven, proposed an
electrowetting display of a rather different type. The pixel was a white
surface, which in quiescence was covered by a thin film of coloured
oil. Application of 20 V caused the film to curl into the pixel corner,
causing a marked change in reflectivity (figure 15). For small pixels the response time was about 10 ms (Hayes & Feenstra 2003).
The reflectivity is much greater than that of LCDs, and approaches that
of paper. In transmission the technique may offer less advantage, but
the application to electronic paper is obvious.
Electrowetting characteristic (after Hayes & Feenstra 2003).
Understanding of electrowetting is incomplete, but some extensive reviews are now available (e.g. Mugele & Baret 2005).
Only certain surfaces demonstrate the theoretically predicted
behaviour, but as the technique acquires commercial importance, it is
likely that more detailed physical models will emerge.
(b) Emissive displays
The
exploitation of emissive effects in solids raises an immediate problem.
If the solid surface is planar and the emitted light is incident on the
interface at an angle greater than sin−1(1/n), n being the refractive index of the solid, the light cannot emerge (figure 16).
Frustrated total reflection.
The loss is serious for many semiconductors used in LEDs. For example, n is 3.66 for GaAs, so the critical angle θc
is only 16°, and only 4 per cent of the energy radiated into a
hemisphere reaches the outside world. Fortunately refractive indices
decrease as energy gaps increase, but even so only 10 per cent of the
radiation from a semiconductor emitting at visible wavelengths would
emerge. It is possible to reduce the loss in an LED by shaping the
surface as a dome or Weierstrasse sphere (Carr 1966), but that would not be simple for a display, with perhaps 106 emitting pixels each less than 1 mm2.
(i) Inorganic solid emitters
The
conversion in a solid of electrical energy into light was first
observed in 1907 by Henry Round, of Marconi, who saw yellow light
emitted from a SiC crystal when he applied a voltage to it (Round 1907).
Round published his discovery in an admittedly obscure journal, and it
attracted little interest. In 1923 Oleg Lossev at the Radio Laboratory
in Nijni-Novgorod, unaware of Round’s work, repeated the experiment, and
published in higher-impact journals (Lossev 1923, 1928).
Diode electroluminescence (EL) is therefore known as the Lossev effect.
LEDs are now universal, and it was natural in the early days of
flat-panel R&D to consider their relevance. LEDs are made with high
density on single-crystal substrates, and the size of the substrates is
rarely more than a few centimetres square. Assembling substrates on a
large area and handling the interconnections would be expensive, and
there has been no serious attempt to fabricate a TV display for sale (figure 17).
Moreover, at a time when subtractive effects were making slow progress,
no efficient blue LED was available. Today a full range of colours is
available, and huge LED displays are seen in many sports stadia.
An early tour-de-force in LED assembly: multi-colour LED flat-panel display with 64×64 picture elements.
A more practicable approach to EL had actually come in 1936, when Destriau (1936)
applied an AC field to a capacitor made with ZnS powder immersed in a
dielectric. His device and the subsequent industrial versions were not
used extensively because of their short life at high brightness.
Lower-brightness panels, made by firing a coating containing ZnS
particles on to a steel plate, were marketed by GTE Sylvania, but their
application was as night-lights, not displays. The modern version of the
Destriau panel, shown in figure 18, is made entirely from thin films, avoiding most of the degradation processes of powders.
Above a critical field, typically 106 V cm−1,
insulating layers of ZnS will show a rapidly increasing current with
voltage increase, and they simultaneously emit light. Generally
manganese centres are incorporated in the ZnS to obtain yellow emission,
but other centres, such as rare earths, can also be used. The problem
is to operate the layer close to breakdown without risking current
runaway. In the AC device shown, only a limited amount of charge is
permitted to pass through the ZnS, and this amount depends on the
capacitance of the dielectric. Clearly the dielectric must not break
down, nor must it demand too high a fraction of the voltage drive. To
ensure that the field in the ZnS is above the threshold required for
light emission ET and the dielectric is kept below its own breakdown field EDB, the inequality
4.7
4.8
- — The dielectric must be thin.
- — The dielectric constant must be high.
- — The breakdown field must be high.
Such
panels were made in 1974 by Sharp, and a few years later by Tektronix.
They each faced the breakdown problems with some success, but the only
reliable efficient phosphor available was ZnS doped with Mn, giving an
excellent yellow, but no good blue or red. A spin-off company from
Tektronix, Planar Systems, did report blue-green emission from SrS:Ce,
but, in practice, only small AC electroluminescent (ACEL) panels were
ever produced in number because of the competition from LCDs.
A
major defect of the ACEL panel, which has excellent appearance, and can
be multiplexed to a high level, is the high voltage of operation.
Special circuits must be manufactured, and these add much to the cost of
the panel. It should be simpler, and cheaper, to drive the panels DC,
but there are severe problems in driving a large-area device near to
breakdown without risking catastrophic failure. A film that is
non-uniform in thickness, composition or geometry will obviously break
down at weak points, either where the layer is abnormally thin, or where
there is field amplification at edges or points. If we construct a
uniform film, there is still the possibility of an S-shaped negative
resistance (figure 19), which leads to current concentration in filaments, with consequent breakdown due to local heating.
It
follows that stable DC thin-film EL over a large area requires a layer
that adds a current-limiting resistance in series, the DC equivalent of
the AC dielectric. This must be done without adding appreciably to the
operating voltage, with a consequent reduction in the already low
efficiency. The problem is not simple.
Successful DC
electroluminescent (DCEL) panels were made, but these relied on phosphor
powders packed closely enough together to give conduction. In addition,
the ZnS:Mn powder had Cu added, and this was often in the form of a
surface-rich layer. In general, however, panels made in this way had low
brightness, low efficiency and short life. A major step forward was
made by Aron Vecht, who had led research on EL at Associated Electrical
Industries, Rugby, but left in 1968 when the company merged with GEC,
and the responsibility for EL passed to the Hirst Research Centre,
Wembley. Vecht and some key members of his team moved to Thames
Polytechnic, later to become the University of Greenwich, and there he
invented an ingenious way of creating stable DCEL panels (Vecht 1970; Vecht & Werring 1970).
The
ZnS particles were coated with Cu, immersed in a polymethylmethacrylate
or nitrocellulose binder, spread over a glass plate previously coated
with a transparent conducting film, and having an Al or Cu electrode
applied to the surface to act as a cathode. The powder concentration is
so high that a conducting path exists between the top cathode and the
transparent conductor.
The emitting region is formed
by passing a current high enough to heat the panel, and over a few hours
the panel impedance increases. The applied voltage is steadily
increased from a few volts to a maximum value, typically 80–100 V, to
maintain the consumed power approximately constant. The electric current
passing through the panel produces a narrow high-resistivity
light-emitting barrier (typically a micrometre thick) near the positive
front electrode film, and it is the gradual formation of this region
that causes the increase in panel impedance. Presumably this region is
formed because Cu diffuses away from the anode (figure 20).
By
1980 Vecht had improved the performance of DCEL panels enough for them
to be considered for consumer applications. When driven at 100 V, they
gave an initial brightness of 300 cd m−2, and had a life to half-brightness of over 15 000 h (Vecht 1982).
Two problems prevented wider adoption. Though the Mn emission spectrum
was broad enough for good reds and greens to be obtained by filtering,
no efficient blue emitter was discovered. Rare earths did give a good
narrow blue line, but the solubility was too low for the emission to be
useful. The other problem was the cost of 100 V driver chips, and this
made both ACEL and DCEL uncompetitive. However, that situation changed
when the potentialities of organic materials for EL became apparent.
(ii) Organic solid emitters
There was little interest in EL in organic materials for some years after Destriau’s work, but in 1953 one French group, led by Bernanose (1955),
and in close collaboration with Destriau, observed ACEL in acridine and
carbazole at voltages up to 2 kV and frequencies of 50 Hz. In 1960
Edward Gurnee and Reet Fernandez of Dow Chemicals showed that a variety
of organic compounds could be used, and found that most light was
emitted when the phosphor was a host containing a doping agent or
activator, and the colour depended on the combination. A typical
phosphor had anthracene as the host, with 0.1 per cent tetracene as the
activator. The EL cell operated off 800 V at a frequency of 2 kHz (Gurnee & Fernandez 1960).
By 1962 Gurnee had added a thin insulating layer in the cell
construction, thereby reducing the operating voltage to 110 V and the
frequency to 60 Hz (Gurnee 1962).
At about the same time we saw the first DCEL organic devices, when Martin Pope4
used a thin crystal of anthracene with NaCl solutions as electrodes and
applied voltages of 400 V or more. It was apparent that the only way of
passing current through a material that was insulating at low voltages
would be to design injecting contacts, and the experiments had limited
success (Pope et al. 1963). Dresner & Goodman (1970a,b)
made some progress in 1970 with a tunnel cathode, but the operating
voltage was still nearly 100 V. The next forward step remained
unacknowledged for several years, being described in a patent by Roger
Partridge, of NPL, in 1975 (Partridge 1975).
He set out clearly the device design for an organic LED (OLED): a
low-work-function cathode, to inject electrons, a high-work-function
anode, to inject holes, and, between them, a luminescent organic film,
which he specified as being a polymer, such as polyvinylcarbazole.
Partridge’s patent is specific on the materials and thicknesses of the
various films, but he did not submit his work for publication until
1981, and his results were not generally available until mid-1983
(Partridge 1983a–d).
The delay was not, of course, Partridge’s responsibility. The
Government department to which NPL reported, the Department for
Industry, wanted to preserve publicity until they could identify
exploitation routes. The secrecy was compounded by entitling the patent
‘radiation sources’, a catch-all phrase (R. H. Partridge 2009, personal
communication). This strategy would have been admirable if the
Department had then used its resources to secure industrial
collaborators, but such efforts as it made were unsuccessful. As a
result, Partridge’s claim for priority has only recently been
acknowledged generally. He was the first to design and make polymer LEDs
(PLEDs) with injecting electrodes, though he did not specify the
advantages of using conjugated polymers, one key to later development.
Before
Partridge’s paper appeared, others had made important contributions.
Significant was the work of Gareth Roberts’ group at Durham University,
which first obtained EL in 1979 from a Langmuir–Blodgett film of
anthracene (Roberts et al. 1979), and in 1982 used a vacuum-deposited film of anthracene less than 1 μm thick, and saw blue light at a voltage of 12 V (Vincett et al. 1982).
The
potential of the field was now apparent, and many groups were working
on EL in small molecules or polymers. Noteworthy was the group at Kyushu
University, which seem to have been the first to obtain results from
layered devices, using cyanine dyes (Era et al. 1985), and then perylene films (Hayashi et al. 1986).
Members of this group, Chihaya Adachi, Shizuo Tokito, Tetsuo Tsutsui
and Shogo Saito, specifically prescribed the structure necessary for
low-voltage efficient OLEDs, shown as a schematic in figure 21.
Early in 1988 they reported three-layer devices emitting blue, green
and orange-red light from anthracene, coronene and perylene respectively
(Adachi et al. 1988).
The
device is made on a glass substrate coated with ITO. ITO has a high
work function, and acts as the anode, injecting holes. The next layer
must transport holes (hole transport layer, HTL), but should also
inhibit the passage of electrons. Next to this layer is the
electroluminor, where the holes and electrons will recombine, and there
is then an electron transport layer (ETL), inhibiting hole passage. The
final metal cathode has to be a low-work-function metal, such as Mg:Ag.
This structure is now fairly standard, though in some designs one
material can perform two functions. The physics underlying the work of
the Kyushu group cannot be faulted, but their embodiment did not perform
well, with a very low efficiency, a drive voltage of 50 V, and a life
of only 5 h.
Adachi had been a PhD student at Kyushu,
and he has continued to make significant contributions to the field at
various laboratories since then, returning to Kyushu 4 years ago. His
early collaborator at Kyushu, Tsutsui, remained there throughout his
career, again making significant contributions until his recent
retirement. Their work has, perhaps, received less acknowledgement than
have the groups linked with industry, but both have certainly been major
forces in OLED development since its inception.
Ching
Tang had worked at Eastman Kodak in the early 1980s on organic EL cells
similar to those developed for ACEL in ZnS, described in the previous
section. He had substituted tetraphenylethylene or similar organics for
the ZnS, held them in a polymer binder, and found that conduction was
possible when a layer of a porphyrinic compound was inserted between the
anode and the organic emitter. The cell operated at quite low voltages,
around 30 V, but the efficiency was well below 1 per cent (Tang 1980).
He then abandoned the binder, and in 1987, with Steven VanSlyke,
published the classic paper on OLEDs, with electrons emitted from an
Mg:Ag cathode into a layer of Alq3, tris(8-hydroxyquinolinato)aluminium, with a layer of diamine acting as the HTL. The Alq3
layer plays two roles, acting as both the electroluminor and the ETL.
Green light was emitted, with an efficiency approaching 0.5 per cent at
5.5 V. The life of the device was 100 h, though the voltage had to be
increased steadily to maintain the light output (Tang & VanSlyke 1987).
This paper has now been cited over 5000 times, and was followed by a
string of publications and patents over the next 20 years. Much of the
progress in the field has been due to the work of them and their
colleagues at Eastman Kodak.
Alq3 has
proved a critical component in most advanced OLEDs since its favourable
properties were shown by Tang and VanSlyke. It is a complex molecule
that can adopt a number of crystalline phases, but, since it is usually
deposited by evaporation, resulting in an amorphous film, control of the
composition is not easy. Moreover, though for many years it was
believed that there was a single isomer, it is now clear that there are
two, one meridional and one facial, shown in figure 22 (Cölle & Brütting 2004).
The characteristic green emission is due to the meridional isomer, and
the facial, often present as a few per cent in amorphous films, emits in
the blue.
Isomers of Alq3: (a) meridional; (b) facial (violet, Al; red, O; blue, N; after Cölle & Brütting 2004).
PLEDs
had remained unnoticed after Partridge left the field. However, a group
at Cambridge University, led by Richard Friend, had become interested
in the possibility of OLEDs, but believed that the Kodak approach had an
inherent problem of structural instability due to recrystallization.
They thought that a better luminor would be a polymer, and, in
particular, a main-chain conjugated polymer. They noted the ease with
which thin films of polymers could be prepared by spin-coating, and in
1990 made simple diodes, with poly(p-phenylene vinylene) (PPV),
a polymer known to show bright green photoluminescence, spun onto
ITO-coated glass, with a cathode of Mg–Ag. The devices emitted green
light at 15 V, albeit at low efficiency, and with a life of a few days (Burroughes et al. 1990; Friend et al. 1990).
Their publication attracted much attention, and the approach was
followed by a number of laboratories. Alan Heeger, later to receive a
Nobel Prize for the discovery of conductive polymers, modified both the
PPV polymer and the cathode, and obtained an efficiency approaching 1
per cent (Braun & Heeger 1991).
There
was early realization that the simple sandwich was unlikely to provide
the environment needed for high conversion efficiency, with equal
numbers of electrons and holes available for recombination in the
polymer luminor. There seemed to be some resistance to designing a cell
with the full gamut of layers shown in figure 21,
though Friend, with colleagues including Adam Brown, Jeremy Burroughes
and Donal Bradley, did show that an ETL of biphenylylbutylphenyl
oxadiazole (butyl PBD) gave improved efficiency (Brown et al. 1992). Adachi had previously shown that butyl PBD performed well as an ETL in blue-emitting OLEDs (Adachi et al. 1990).
It was also appreciated that acceptable device lifetime could be
achieved only if the cells were encapsulated to protect them from the
effects of oxygen and moisture. This was particularly necessary if Ca
was used as a cathode, capitalizing on its low work function. There
followed a period of steady progress in both efficiency and lifetime,
and the use of a variety of luminors to give a full range of colours.
By
the turn of the century both organic (OLED) and polymer (PLED) devices
had made such progress that commercial exploitation appeared likely.
Efficiencies of both types and of all colours were well above 1
per cent, and some designs were giving life at constant current of over
10 000 h. There was some concern that the drive voltage almost doubled
during that period, and this would certainly complicate incorporation in
multi-pixel panels. Blue devices were still lagging in performance and
lifetime, but it was thought that this problem would yield to increased
research. Commercial interest of Kodak had always been high, and in the
UK Cambridge University spun out Cambridge Display Technology (CDT) to
exploit PLEDs. Many of the world’s electronic companies were beginning
to become involved.
Efficiency lagged behind desired specifications for brightness. The refractive loss analysed at the beginning of §4b
was not easy to overcome, but several remedies were tried. Almost half
the loss came because most of the light emitted away from the anode was
lost. Composite metal–dielectric layers could reduce this loss. The
plane glass surfaces could be etched to give scattering, or low-index
aerogels interposed (Smith & Hilsum 2003).
Microlenses could be attached to the surface, though this would prove
problematic for small pixels on large areas. One remedy that attracted
considerable effort was the concept of microcavities. This entails
forming Fabry–Perot etalons at each pixel by interposing a dielectric
mirror stack between the glass and the ITO. This should give
directionality, and in principle enable most of the radiation that is
produced by recombination to emerge. There are two obvious problems. One
is the need for different mirror stacks for each pixel colour. The
second is the need for an appreciable angle of view (Peng et al. 2003; Shiga et al. 2003).
Though
the refractive loss is serious, it does not account for the relatively
low internal efficiency. The excitons formed in the process of
recombination take either a singlet or a triplet spin configuration, and
simple theory indicates that the ratio of these states is 1 : 3.
Unfortunately the more numerous triplet states are non-radiative, so
that the maximum internal EL efficiency would be 25 per cent. Since an
isotropic emitter with the refractive index of PPV or Alq3
could emit into the outside world only 25 per cent of the photons
produced, the maximum external efficiency could not exceed 6 per cent.
Measurements confirmed that the singlet fractions in Alq3 were near 20 per cent (Segal et al. 2003).
The situation was not as serious for polymers, for early measurements
showed that the singlet/triplet ratio could be as high as 50 per cent,
and detailed theory indicated that the actual ratio would depend on the
polymer, and in particular on the energy gap (Wilson et al. 2001).
There was, nevertheless, good reason to explore methods for reducing or
eliminating non-radiative recombination for both OLEDs and PLEDs. Two
approaches proved effective.
The first proposal for
reducing the effect of the non-radiative triplet states came from
Herbert Börner and his colleagues at the Philips Laboratory in Aachen.
They filed a patent in 1994 suggesting that the OLED should incorporate a
phosphor having emitting triplet states at energy levels available for
transfer of energy from the luminor non-radiative triplet states (Börner et al. 1995).
Though Börner clearly appreciated that this should give an improvement
in performance, he gave no details of any gain he actually achieved, and
there was no subsequent publication. The concept of exploiting
phosphorescence was more clearly stated by the group at Princeton
University led by Steve Forrest and including Mark Thompson and Paul
Burrows. The group had previously shown that organic devices suffered
severe degradation if exposed to the atmosphere during life, and
demonstrated the improvements on encapsulation (Burrows et al. 1994).
They now played the major role in establishing the science of
phosphorescent doped OLEDs, or PHOLEDs, and filed a number of patents
and published many papers from 1997 onwards (Thompson et al. 1997; Baldo et al. 1998; Baldo et al. 2000).
Thompson had moved to the University of Southern California (USC) in
1995, but he still worked closely with Forrest’s group in the search for
efficient OLEDs, and their results were exploited via a collaboration
with United Displays Corporation (UDC), a company founded in Ewing, near
Princeton, in 1994. UDC has grown to be one of the major suppliers of
materials to the display industry. The invention of phosphorescent
doping played a major role in achieving commercially viable OLED
efficiencies, and world efforts are well summarized in Holder et al. (2005).
The
second initiative was relevant to PLEDs. In the 1980s, Dow Chemical
developed and trademarked Starburst dendrimers, a novel type of
three-dimensional polymer with hollow interiors and surfaces that can be
closely packed with a variety of molecules (figure 23).
The
materials are extremely versatile, spanning many structural and
functional applications, and if the right molecules are incorporated,
they are efficient phosphors. The team that decided that dendrimers
could make efficient PHOLEDs was led by Paul Burn and Ifor Samuel. Burn
had earlier been a member of the Cambridge University PLED team, but
moved in 1992 to the University of Oxford, hoping to develop conducting
dendrimers. Samuel had set up his own group at the University of Durham
in 1995, with an interest in light-emitting dendrimers. Together they
showed that dendrimers could be designed to give efficient blue, red and
green PLEDs, and founded Opsys in 1997 to exploit their ideas (Samuel et al. 1998; Ma et al. 2002; Burn et al. 2007).
CDT bought Opsys in 2002, appreciating that dendrimers could be
effective in improving the performance of their PLEDs, and could be
deposited by spin-coating or ink-jet printing, Some of their current
commercial PLEDs incorporate dendrimers.
Another
pioneer in this application of dendrimers was Jean Fréchet, who had
earlier developed photoresists for integrated circuits, and later used
dendrimers for drug delivery and tissue scaffolds (Gilat et al. 1999; Adronov & Fréchet 2000).
A
development that may well affect the future design of OLEDs came from
the Technical University of Dresden, where Karl Leo, Jan Blochwitz and
Martin Pfeiffer had been considering the analogy of OLEDs with inorganic
LEDs. They thought that the operating voltage of OLEDs was
unnecessarily high because of the two blocking layers inherent in all
previous designs. Their aim was to make an organic p-i-n diode, with
holes injected from the anode into an intrinsic region where they would
combine with electrons injected from the cathode. Obviously this
entailed controlled p- and n-doping of the appropriate layers. They
accomplished this not by adding donor and acceptor elements, but by
molecular doping, and in a series of publications from 1998 to 2003
reported that controlled doping was possible (Pfeiffer et al. 1998; Werner et al. 2003).
A spin-out company, Novaled, was founded in 2003, and this has now
become a major player in the field, specializing in PIN-OLEDs. Since the
improvement comes from the contact layers, advances in either
dendrimers or phosphorescents can be exploited in the intrinsic region
of the diode, so Novaled works on both polymer and small-molecule LEDs.
The current state of the art is shown in table 4.
Some caution is appropriate here. The result from Wei-Sheng Huang,
Jiann Lin and Hong-Cheu Lin, in Taiwan, though a striking achievement,
may not indicate commercial potential, because life data are not
provided (Huang et al. 2008).
On the other hand, information from companies may not represent the
true state of their art, because commercial confidentiality may restrain
them from revealing the full extent of their capabilities. The
commercial competition is very fierce, and as a result it is difficult
to give the composition of the materials favoured for devices, for
either POLEDs or OLEDs. In any case, today’s materials are likely to be
supplanted in the near future by new formulations.
Table 4.
Performance data for commercial OLEDs and PLEDs.
| efficiency | efficiency | efficiency | life (h) | CIE | |
|---|---|---|---|---|---|
| (Cd A−1) | (%) | (l W−1) | at 1000 cd m−2 | coordinates | |
| red | |||||
| CDT | 30 | more than 260 000 | 0.63, 0.37 | ||
| UDC | 28 | 20 | 330 000 | 0.64, 0.36 | |
| Kodak | 12 | 8.5 | more than 70 000 | more than 0.66 | |
| Novaled | 12 | 12 | 10.6 | more than 500 000 | 0.68, 0.31 |
| blue | |||||
| CDT | 9.4 | 10 000 | 0.14, 0.14 | ||
| UDC | 9 | 4 | 0.14, 0.13 | ||
| Kodak | 8 | 6.7 | 10 000 | 0.14, 0.15 | |
| Novaled | 9 | 5.5 | 8.4 | 0.14, 0.23 | |
| green | |||||
| CDT | 21 | 110 000 | 0.29, 0.64 | ||
| UDC | 67 | 19 | 250 000 | 0.38, 0.59 | |
| Kodak | 30 | 8.7 | 65 000 | 0.25, 0.66 | |
| Novaled | 95 | 23 | 116 | 0.31, 0.66 | |
| Taiwan group | 61 | 17.6 | 32 | 0.29, 0.62 |
There
remain several serious hurdles to overcome in the application of these
devices. Spin-coating was thought to give PLED manufacture an advantage
over OLED sublimation, but colour patterning is not simple, and several
companies have concentrated on ink-jet printing. Similarly, mass
production of large areas by sublimation is not cheap, and UDC has
pioneered organic vapour-phase deposition (OVPD), as used for many years
for compound semiconductors (Forrest 2004).
They showed that this gave good results for PHOLEDs, and worked with
Aixtron, which now sell OVPD commercial production equipment. Kodak, for
their part, have developed vapour injection source technology, which
they claim is suitable for large-area production, and is more economic
than sublimation (Kyoshi 2006).
One
important recent development that could either improve the future
prospects of flat-panel organic displays or act as a considerable
diversion of resources is the realization that organic diodes could be
efficient domestic and commercial sources of illumination. Tungsten
filament lamps have a light production efficiency of less than 20 l W−1, or 3 per cent, and fluorescent tubes give at best 100 l W−1
(15%). Since lighting constitutes a high proportion of a country’s
energy needs, government interest in this field has grown rapidly, and
has been accompanied by generous financial support. It is no surprise,
then, to see the major display players establish lighting divisions. The
two applications have much in common, but there are significant
differences. The most obvious is colour, where the illumination need is
for a white, or ‘warm white’ panel, and here progress has been rapid.
Many laboratories have reported white LEDs with external quantum
efficiencies of 15 per cent, and the University of Dresden obtained an
efficiency of 90 l W−1, with the prospect of 124 l W−1 if measures were taken to reduce the total internal reflection (Reineke et al. 2009).
(iii) Gas plasma discharge
Light
emission from gas discharges is over 200 years old, and the use of
inert gases to give specific colours dates back over 100 years. The
first attempt to exploit the effect in a display was made 40 years ago,
when Don Bitzer, Gene Slottow and their colleagues at the University of
Illinois proposed a display panel in which the discharge was restricted
to local areas by a glass sheet containing small holes. This was
interposed between two more sheets on which anode and cathode electrodes
had been deposited as row and column electrodes (Bitzer & Slottow 1966; Bitzer et al. 1966).
The first panels used neon, but soon the standard design used UV
emission from xenon to excite phosphors giving red, blue and green
pixels.
In modern panels the column electrodes are
deposited as metal stripes on a glass substrate, and above them are a
series of parallel ridges, dividing the display area into a number of
narrow troughs. The sides of the troughs are coated with the phosphors (figure 24). The row electrodes are ITO stripes on the glass sheet that forms the top of the panel. The I–V
characteristic of a gas discharge is an S-shaped negative resistance,
and the working point is determined by the resistive load line. It is
possible to energize the pixel by a certain voltage, and sustain it by a
lower voltage, so often there are two sets of row electrodes. The
driving voltage is 2–300 V.
Though
the ridges reduce optical crosstalk, a fault of the original design,
electrical charges on a pixel can influence neighbouring pixels in the
same trough. This can be reduced by shaping the row electrodes, or, in
the limit, by isolating each pixel cell. Nevertheless, it has proved
very difficult to reduce the pixel dimensions below 0.5 mm, so that a
high-resolution display smaller than 30 inch diameter is impracticable.
On the other hand, very large displays are possible, and in the early
period of flat-panel development, the large-panel market was dominated
by plasma panels.
A brightness of 1000 cd m−2
was readily available, with excellent colour rendering, a wide viewing
angle and a rapid response time. However, the efficiency of light
production was low, so the power consumption was as high as for a CRT.
Moreover, the need for high voltages made the electronics quite
expensive. For some years plasma panels had a reputation for short life.
The phosphors degraded, and the blue material was particularly bad. A
life of 10 000 h was not uncommon, though manufacturers today would
claim a much improved performance. One other fault has proved more
difficult to overcome. This is the tendency for the display to hold the
memory of a persistent bright image.
Commercial
interest in plasma display panels (PDPs) was high from 1990 onwards,
with first Fujitsu in the lead and then competition entering from many
Far Eastern companies. By 2004 normal economics reduced those interested
to Fujitsu–Hitachi, Pioneer, LG, Samsung and Panasonic, with the last
two holding two-thirds of the market. In 2008 Panasonic claimed to have
made the largest flat-panel display in the world, with a 150 inch
diagonal (Ashida et al. 2008).
The PDP market grew from 2000, as the lure of the CRT was fading, and
the advantages of the LCD had yet to become apparent. The superb
performance of the panels convinced many that the device had a promising
future, but the manufacturing problems were not common knowledge. In
fact, it was a difficult device to make in large quantity, and the cost
of production was high. The emergence of rival display effects persuaded
Pioneer and Fujitsu to abandon plasma panels early in 2009, and the
fact that the total market is reducing may well mean that plasma panels
will become a niche commodity.
(iv) Flat cathode ray tubes
The
engineering world had been very conscious in the mid-1900s of the
disadvantages of the CRT display, and it was not sitting idly by waiting
for an alternative to appear. The main problem was the cumbersome
shape, which was a great hindrance in moving to a larger screen size.
The designers of a flat replacement had to face at least two severe
problems. The first was to arrange the gun, itself quite long, within
the depth of the tube, and then deflect the beam through two right
angles without losing focus. The activation of the phosphors available
at that time required almost 10 kV, and the beam current needed for a 40
inch screen would approach a milliampere. Such beams are difficult to
deflect. The second problem is the vacuum within the tube. Safe practice
demands that the thickness of the glass face plate must be at least
1/15 of the unsupported width, so that a 30 inch diagonal flat panel
would need two plates over an inch thick and would weigh nearly 50 kg.
One solution is to use supports bridging the gap, but these may
interfere with the picture.
One prolific inventor who
attempted to find a solution was Dennis Gabor, a Hungarian who first
worked for Siemens in Germany, but came to the UK in 1934, and later
earned a Nobel Prize for holography. His interest in CRTs was shown by a
number of patents, including some on tubes that had ‘relatively small
depth’, but had an elliptical cross section to accommodate the electron
gun (Gabor 1955, 1956, 1957, 1959). He did not claim a truly flat tube until 1962 (Gabor 1962; Gabor et al. 1968).
The
other early inventor was Ross Aiken, who started his research in 1951,
but waited to file patents until he was sure he had the best design, and
could get financial support for the development. He secured that
support from Kaiser Aircraft in 1953 (Aiken 1984), but when he filed his patent in December (Aiken 1953),
he learned that Gabor had anticipated him. There followed a patent
battle, which was resolved with Aiken having the US rights, and Gabor
the rights in the UK. Neither tube was ever produced on a commercial
basis.
Ten years later RCA mounted a serious programme
aimed at a 40 inch flat-panel colour CRT. Two approaches emerged. In
that led by John Endriz, an x-array of 500 line cathodes fed electrons
into 480 vanes of secondary emission multipliers (Cosentino et al. 1979; Endriz et al. 1979).
The other team, including Tom Credelle and Tom Stanley, favoured a
normal thermionic cathode injecting electrons into a series of box
guides at 90° to the line of the gun. These boxes, which acted as
supports between the two display faces, contained a series of plates to
deflect the electrons through a further 90° on to the phosphor (Stanley 1975, 1977; Siekanowicz et al. 1976; Credelle 1977).
Other attempts were made by Texas Instruments (Scott 1977) and Siemens (Graefelfing et al. 1968),
but by 2000 it was generally appreciated that the problems of
brightness, resolution and display size could not be met in displays
that would be competitive commercially. Any solution would have to come
from a design that bore little resemblance to the CRT. No matter how
clever inventors were in twisting and bending electron beams, the
obvious way forward was to make a cathode that was itself a plane, and
that could be arranged parallel to the phosphor screen. It would be
difficult to do this with a thermionic cathode array, since the screen
would itself get hot, not a desirable attribute.
Ken
Shoulders and Louis Heynick, working at the Stanford Research Institute
(SRI), were the first to appreciate that the answer could come from cold
cathodes, with the field required to extract electrons reduced by
shaping the cathode into a point. They made small cathodes as a clump of
needles of molybdenum or tungsten, formed by coating the surface with
aluminium, and then heating until the Al melted (Shoulders & Heynick
1966a,b).
They observed that surface tension then caused material to migrate into
the needle shapes, but candidly admitted that they did not understand
the mechanism. A few years later Capp Spindt collaborated with Heynick
to invent a more practicable manufacturing method (Spindt & Heynick 1970). They evaporated Mo on to a plane surface, and used an ingenious shadowing technique to form minute cones (figure 25).
Their
success caused many others to join in the search for a flat CRT, and
the Spindt cathode became the standard basis on which flat CRTs were
developed (Spindt et al. 1991).
A number of companies were formed to make devices, but their success
was limited, and most saw a period of serious investment before it was
realized that the technical problems militated against commercial
success. Monochrome tubes were made, in sizes up to 12 inch, and colour
tubes were described as producing a picture equal to that of a normal
CRT. However, it was difficult to get sufficient sustained emission for a
screen brightness above 500 cd m−2, and production costs
were high. The major player was Candover, effectively a spin-off from
SRI founded in 1990, which attracted funding of over $300 million. In
2004 they sold their technology to Canon for $10 million, and abandoned
the field. PixTech, a French spin-off from LETI in Grenoble, with a US
subsidiary, was founded in 1992, and they also tried to market displays
using their own version of Spindt cathodes. They had a limited success,
and demonstrated 15 inch colour tubes at exhibitions in 1999, but there
has been little activity in the last five years.
Among
the companies interested in cold cathodes was the UK General Electric
Company (GEC). Their main potential application was microwave tubes, but
they were conscious of the display possibilities. A group at the Hirst
Research Centre led by Rosemary Lee first concentrated on making
microtips from silicon, since they thought that the production
techniques were simpler than those for Spindt cathodes. They did succeed
in making efficient cathodes that were stable (Lee et al. 1989),
but it was clear that this technique could not give cheap large-area
sources. However, in 1987 the group also invented a planar version,
which was potentially suitable for large-area metal cathodes (Lee &
Cade 1987, 1988; figure 26).
There
was clear merit in a surface structure that attracted electrons
parallel to the surface and scattered a proportion of those incident on
the anode towards a second anode on the other side of the panel. The
surface emitting display (SED) could be fabricated by the techniques
perfected for chips, albeit on a larger scale. Many variants of the
planar cathode were invented subsequently, with the lead taken by Canon
and Futaba. Futaba, with a history of manufacturing small vacuum
fluorescent displays, showed that carbon nanotubes deposited on the
surface gave a larger and more stable current than metal edges (Itoh et al. 2007).
Canon also made steady progress. They teamed first with Toshiba, and,
after this collaboration ended because of a legal dispute with Applied
Nanotech, with Hitachi and Matsushita. At regular intervals after 1998
it was announced that commercial production was imminent, with the press
releases after 2006 given weight by the impressive panels shown at
exhibitions. In 2006 Canon demonstrated a panel with a brightness of
450 cd m−2, a contrast ratio of 50 000 : 1 and a response
time less than 1 ms. However, 3 years later there is still no sign of
commercial production of TV panels. Obviously the SED is not easy and
cheap to make.
5. Components for addressing
In
§2 I discussed the need for an electrical nonlinear component to be
integrated on the panel. This need came to the fore in the 1970s, when
the potential of LCD panels was becoming apparent, but there was no
candidate component for addressing. It must be admitted that this caught
the international semiconductor expertise off balance, in that almost
all effort had been directed at growing large perfect single crystals,
and then depositing thin layers of controlled purity epitaxially on the
crystal slices. Materials experts felt that glass substrates were
incompatible with silicon, germanium and the III–V compounds, and those
semiconducting compounds that could be deposited on glass, like cadmium
selenide and zinc sulphide, would never give stable performance for the
several years that companies expected. Instead they turned to
metal–insulator–metal sandwich diodes, with combinations like Ta/Ta2O5/Cr, or glass switches, made of combinations incorporating chalcogenides, such as GeAsTe (Ovshinsky & Fritzsche 1973; Adler et al. 1980).
Though some of the research on these components was generously funded,
and enjoyed collaboration with internationally renowned physicists,
stability, reproducibility and long life were never obtained.
The
solution came from the semiconductor workhorse, silicon, but in an
unusual form. In the 1960s, a group led by Henry Sterling at the
Standard Telecommunications Laboratory at Harlow had shown that the
glassy form of Si, normally considered an insulator, could be made
partially conducting if thin films were deposited in a hydrogen
atmosphere (Chittick et al. 1969).
The research was continued at Dundee University by Walter Spear and
Peter LeComber, and in 1975 they showed that the conductivity could be
controlled by the addition of the same impurities as were used for
doping crystalline Si (Spear & LeComber 1975).
This discovery inspired worldwide efforts that led to the first
practical solar cells. At that time, the interest in renewal energy was
beginning to gain momentum, and it was not surprising that Spear saw
that his priority should be a focused strategy aimed at keeping his
research group at the head of the international a-Si solar cell effort,
leading in the basic physics.
The Displays Group at
the Royal Signals and Radar Establishment (RSRE), Malvern, had a
different view. They saw a UK laboratory in possession of the key to
addressing displays. They were sure, following instinct rather than
science, that a-Si TFTs would work with LCDs, but they lacked the
facilities for making the devices, no easy task. Persuading Dundee to
collaborate, while their eye was on a different ball, took many months.
In any case, Spear had reservations about collaboration with
non-academic organizations (Adams 2009).
It was LeComber, without Spear’s blessing, and possibly without his
knowledge, who eventually made the original a-Si TFTs, and supplied the
first batch to Malvern. RSRE switched LCD pixels with them on Xmas Eve,
1978. The structure of the TFT and its characteristic are shown in figures 27 and 28 (P. G. LeComber 1978, personal communication). Later work at Dundee and Malvern confirmed the potential of a-Si for displays (LeComber et al. 1979; Snell et al. 1981).
Electrical characteristics of the first a-Si TFT, made at Dundee University.
Neither
group was able to gain financially from this development, because the
original Weimer patent covered the TFT as a device, and did not specify
any particular semiconductor. The two groups consulted patent experts,
but were advised that an attempt to license the invention would
certainly involve legal action, which they were unlikely to win. So this
discovery, which is still widely used worldwide, brought no direct
economic benefit to the UK.
Though the basis for
progress was now established, there were many hills to climb before
industry was convinced of the virtues of this new approach. One major
advantage was the use of silicon, for industry knew that this
semiconductor was stable, unlike the sulphides and selenides previously
on offer. This conviction was not severely undermined by reports of
changes in solar cell performance during prolonged exposure to sunlight.
The cause of this effect, first reported by Staebler & Wronski (1977),
is still obscure, and it was not highlighted during the early days of
a-Si addressing. Indeed, the protagonists of more exotic materials like
CdSe for TFTs could well have argued that the playing field was not
level, for they were criticized for changes in device properties during
life that were no worse than those for a-Si (Weimer 1962a; Lee et al. 1992; Brody 1996).
However, fair play was never a feature of displays commercialization,
and it is true that much research and huge funding ensured that
stability was eventually obtained, at least in a-Si arrays for LCDs.
Nevertheless, it was many years after the first pixel was switched at
Malvern before large arrays could be produced in industry with a high
yield.
Amorphous silicon has proved essential in the
progress of flat-panel displays, but its semiconductor properties are
undistinguished. In particular, the field effect mobility is rarely as
large as 1 cm2 V−1 s−1, a thousand
times lower than that of crystalline Si. This means that the material
can be used only for the addressing switches. The display requires
drivers and data control circuitry as well, and if a-Si is used for the
switching, standard chips have to be fixed to the panel. A dream of the
early workers was that a semiconductor would be found that would have a
high enough mobility to accomplish all the necessary operations. The
mobility needed was at least 10 cm2 V−1 s−1.
The
answer came from standard chip production, where polycrystalline Si
layers were routinely deposited by chemical vapour deposition (CVD) of
silane, though the deposition temperatures were well above the softening
point of glass. However, it was discovered by Doug Meakin and Piero
Migliorato at GEC, Wembley, in collaboration with Nico Economou of the
University of Thessaloniki, that lowering the chamber pressure from the
standard 26.6 Pa to below 1.3 Pa allowed growth at under 600°C of
polycrystalline Si films with a mobility of 10 cm2 V−1 s−1 (Meakin et al. 1987).
This triggered other laboratories to work on poly-Si, with an emphasis
on alternative production methods. The favoured method was laser
annealing of a-Si, and mobilities gradually increased. The leading
companies were, and still are, Seiko–Epson (Jiroku et al. 2004) and Sony (Sekiya et al. 1994), and both have reported poly-Si TFTs with mobilities of several hundred cm2 V−1 s−1. Seiko was using the TFTs for integrated drivers. The two companies may well combine their display activities soon.
Though
the electrical properties of poly-Si are surprisingly good for a film
containing many grain boundaries, the production process is costly, and
has not been used for large displays. Alternatives are being sought.
Research has been reported on several semiconducting oxides, including
zinc oxide and indium oxide, each of which can have field mobilities
much greater than that of a-Si (Levy et al. 2008; Park et al. 2008; Wang et al. 2008).
In addition, progress on semiconducting organics for LEDs has led to
interest in fully organic back planes, and a number of organic TFTs have
been reported. CDT, using 6,13-bis(triisopropylsilylethynyl) (TIPS)
pentacene, has made TFTs with a gate width of 50 μm and mobility up to
1.8 cm2 V−1 s−1, with little reduction for gates as narrow as 5 μm,
demonstrating the low contact resistance, a key point in organic TFT
research. CDT has also obtained, using a proprietary material from
Merck, a TFT with mobility above 3 cm2 V−1 s−1 (J. H. Burroughes 2009, personal communication).
Though
early work aimed at replacing a-Si TFTs, and used glass substrates,
more recently there has been emphasis on solution or ink-jet printing,
and deposition on plastic substrates for flexible displays. An early
leader was Plastic Logic, spun out from the Cavendish Laboratory, which
has made TFTs by ink-jet printing on plastic substrates with mobilities
approaching 1 cm2 V−1 s−1. Another
forward step has been the formulation by CDT, in collaboration with
Silvaco Data Systems, of an equivalent to the SPICE simulator for
electronic device automation. There is little information on the life of
organic or polymer TFTs, though it would be surprising if this were as
good as for a-Si or poly-Si, particularly for devices on plastic
substrates.
6. Applications
The
spectrum of modern uses of flat-panel displays ranges in size from
microdisplays, requiring optical assistance to view them, to the huge
displays now common in sports stadia, and it encompasses many types of
embodiment. Though there may well be significant differences in the
types of visual effect used for some applications, I will here
concentrate on a few applications that demonstrate the breadth of the
technology.
It should not be forgotten that
even a simple flat panel may yield a considerable advantage in cost and
complexity over alternative ways of presenting data. For example, long
before active matrix techniques were available, automobile manufacturers
were considering incorporation of LCDs or EL panels into the driver’s
display. Not only was the thickness of the display reduced from 60 to
12 mm, but the number of parts and processes came down from 430 to 35.
(a) TV panels
There
were two trends in the consumer market after 2000 that resulted in the
demise of CRT TV sets. The first was the realization by computer users
that a flat panel gave them extra desk top space. They were willing to
pay a considerable premium in price for this privilege. The second was
the desire of households worldwide to view a bigger TV picture. There
was absolutely no fault with the CRT picture—it was bright, sharp, could
be viewed at all angles, and was inexpensive. Certainly the brightness
and colour purity faded with time, but many households were happy to
keep the same set for ten years or more. The change happened very
rapidly. It was, of course, made possible by the improved skills of the
manufacturers, which learned how to make a-Si TFTs with high yield over
large areas, illustrated in figure 29.
The growth in manufacturing capability of active-matrix LCD panels.
The
results on sales were surprising. In the UK, Dixons, then the largest
retailers, reported that in 2004 nearly 90 per cent of the TV sets sold
were CRTs. In 2005 the percentage had dropped to 20 per cent, and in
2006 had dropped to 5 per cent. Dixons stopped selling CRT TVs the next
year, and were shortly followed by other retailers. However, CRTs still
retain niche markets, particularly when high resolution is required.
The
move towards larger TVs had benefited plasma panels (PPs), which were
previously reserved for specialist applications, but after 2000 they
dominated the flat-panel TV market for screens larger than 40 inch. By
2006 the available PP size was just over 100 inch, though larger panels
had been demonstrated (figure 30). Very large LCs were, however, also becoming available.
A 150 inch Panasonic plasma panel TV.
The
two technologies now compete aggressively in the market for large TVs,
with each having protagonists in their views on the excellence of the
picture. PPs certainly had an early edge here, but LCD manufacturers
have made many improvements in back lighting, power consumption and
angle of view, so that there is now little to choose between them. The
current market position is shown in figure 31.
World sales of flat panels, 1986–2008.
PPs
will face great pressure from LCDs and OLEDs in the future. LCD
manufacturers are investing heavily in increasing capacity and reducing
costs by standardizing on huge glass areas, and the sheet of glass from
which displays are cut is now Generation 10, 2880×3130 mm (figure 32).
Corning plans to invest over $400 million in a plant for Sharp at Sakai
City. Samsung does not intend to lag. Their plans are for a plant using
Generation 11 glass, 3000×3320 mm, which would give six 72 inch LC
panels from one sheet.
Generation 10 glass for displays.
The
outlook for OLEDs in TV is not certain. Though impressive OLED panels
have been shown in exhibitions for a number of years, the UK consumer
had to be patient until early in 2009 before they could buy one. Sony
then announced a screen 11 inch wide with excellent picture quality. The
price, however, illustrated the manufacturing problems. The first
announcement quoted a price near £3500, though this was quickly followed
by offers to import panels from the USA at half that price.
It
is seductive to present any light-emitting technology as simpler than,
for example, an LCD, because no additional source of illumination is
necessary. The counter argument is that LCDs separate the problems, so
they can all be attacked individually. The progress in LCD picture
quality attests to the validity of that argument, for much of the recent
research has concentrated on back lights, angle of view, power
consumption economies and other factors that have had lower priority
while the technology achieved its leading market share.
Many
of the OLED production problems arise from the interdependence of many
of the display’s functions. The panel must be able to show the three
primary colours, and the red and the green emitters are eminently
satisfactory. The blue, on the other hand, has too short a life. This
means that the colour balance of the picture will change during its use.
One method of preserving colour balance is to use red, green and blue
phosphors excited by a deep blue or UV source, but here it is that
source which is short-lived.
Another
problem is the addressing. The separation of functions in the LCD means
that the a-Si TFTs are just switches, and they carry a minimum current.
In OLED panels, the current energizing the emitters has to pass through
the TFTs, and they therefore are driven hard throughout their life. Life
changes due to trapping effects in a-Si are of minimal importance in
LCDs, but not in OLEDs. As a result, a number of OLED manufacturers have
turned to poly-Si, which is more stable, but more difficult to make. In
addition, the need to pass current to pixels in the middle of the panel
throws an additional load on the quality of the transparent conductor.
It is not then surprising that the heralded arrival of OLED TV has been
delayed, though the penetration into less demanding markets has
continued apace.
(b) Flexible displays
In
the early days of flat panels, there was some concern about the glass
substrates. The risk of breakage was, perhaps, emphasized, and there
were some experiments with plastic sheets. In time, however, glass
became the standard, and there it has remained. However, at some stage
defence users again became conscious of their dependence on an
admittedly fragile component, and began funding research on plastic
substrates. They took this initiative further, asking for a display that
was flexible, with a vision of field maps that were portable, and could
be rolled up. In fact, a form of flexible display had been available to
the consumer for many years, a temperature indicator containing
cholesteric LC. These were marketed widely, and since they were cheap,
the customer rarely complained when their ability to read the
temperature lapsed after some months. That, however, was symptomatic of
the problem of using plastics. Their ability as a moisture barrier is
far less satisfactory than that of glass, whose main weakness in that
regard is the adhesive seal between the two sheets. The requirement for a
display substrate is a water vapour transmittance less than 10−6 g m−2 d−1 and an oxygen transmittance less than 10−5 ml m−2 d−1, whereas the best common plastics, PET and PEN, are several orders of magnitude worse than this.
The
demand for moisture and oxygen barrier coatings has attracted much
research activity, for the potential applications are broader than just
flexible displays. Included in these were solar cells, which obviously
have to be exposed continuously to the weather. Many companies have
undertaken research, including 3M, Vitex, Dupont and Batelle, and many
government and university laboratories have also been involved.
The
general presumption is that no single material will meet the harsh
specification, so most recipes involve a series of organic–inorganic
layers, sometimes with a thin metal film as the inorganic layer.
Diamond, hard carbon, has been proposed by a number of research workers.
The actual measurement of moisture penetration is not easy. The
standard technique has a limit of 0.005 g m−2 d−1,
well above the requirement. As a result it is difficult to assess
whether the films available commercially are good enough, since they
usually just quote performance as better than the standard limit. Early
in 2008, the Singapore Institute of Materials Research and Engineering
(IMRE) announced that they had developed a film that was 1000 times more
effective than any technology on the market. This had some credibility,
since IMRE had previously invented test sensors that claimed to be
capable of measuring water vapour permeability rates of 10−7 g m−2 d−1. The claim was that incorporating Al2O3 nanoparticulate matter into a stack of layers of acrylic polymer, ITO, Al2O3
and ITO on a polycarbonate film gave very low permeation, and a series
of three stacks met the device requirements. Titanium nanoparticles
could be used together with or instead of aluminium oxide (Ramadas & Chua 2006).
Nevertheless, over a year later there has been no journal publication,
and no independent verification, so judgement must be deferred.
It
seems unlikely that all the production and life problems of making
large flexible video displays will be solved swiftly. Smaller displays,
which give more freedom to the instrument designer because the display
can be conformable, should be available in the near future.
(c) Electronic paper and electronic books
An
electronic book, now generally known as an E-book or Reader, is the
digital equivalent of a printed book, in the simplest versions capable
of presenting monochrome text one page at a time, with page changing
taking a second or two. One of the first people to try to capitalize on
the advantages of an electronic book was Robert Maxwell, who approached
the author at RSRE, Malvern, in 1981, with a project to make a portable
LCD capable of reproducing a page of text. Maxwell was then Chairman of
the British Printing and Publishing Corporation, which owned several UK
national newspapers, but his interest stemmed from his ownership of
Pergamon Press Ltd, which published many legal and medical tomes. These
were expensive to produce, and new editions were needed every year. He
thought that an electronic memory and display was the way forward, and
was prepared to finance generously an R&D project. He stated that
money was no real object. The concept was clearly ahead of its time,
since active matrix techniques were still in their infancy, and adequate
data storage was not then available within the dimensions of a
textbook. The project had to be deferred, probably fortunately, in view
of his subsequent history, but the concept was certainly valid.
It
is simplest to think of an E-book as a Reader in which the visual
effect is provided by a flat or flexible electronic display that rivals
the effect of print on paper, and so is called electronic paper. A
number of versions of the E-book are now on the market. The input data
come from inbuilt memory, or via a wireless connection. The screen size
in early versions approximates to a paperback book, though recent models
are as large as A4 paper. Some designs are based on plastic films,
largely to reduce the weight and make the device robust. One important
feature of the E-book is its portability, so that the power must be
supplied by a rechargeable battery. This militates against emissive
visual effects, though back lights can help for viewing in poor light.
There is obviously a premium on high reflectivity and high contrast.
The type of electronic paper that is most widely used exploits electrophoresis (EP). This effect was discussed previously, in §4a(ii),
and there it was explained that the early technology was found wanting
because the pigment particles clumped, settled under gravity, and left
stains on the electrodes. Modern EP cells reduce these problems by
either microencapsulating the pigment particles in tiny spheres within
the carrier fluid, or by dividing the panel volume into segments by
barriers between the surfaces.
The pioneering research on microencapsulation was done by E Ink, which collaborated with research scientists at MIT (Jacobson et al. 1997; Comiskey et al. 1998). They also worked closely with Philips, at Heerlen and Kobe, and the structure they devised is shown in figure 33 (Henzen et al. 2004). The main differences from the classic form shown in figure 14
are the encapsulation and the use of two sets of pigmented particles,
one set black and negatively charged, the other white and positive. The
fluid within the sphere is clear. In principle, the positive particles
could be coloured, or there could be micrometre-sized colour filters on
the front surface. Commercial e-paper has a reflectivity of 40 per cent
and a contrast ratio of 7 : 1. White paper has a reflectivity of 85
per cent.
Operation
of an EP display requires long positive and negative pulses, and this
complicates addressing. The response time is set by the transit time of
the pigment particles, and for E Ink paper this is just less than a
second a page if grey scales are to be shown. The paper is used in the
Sony Reader and the Kindle, which both use glass substrates. There has
been some concern about the use of glass substrates in the larger
Readers, which could be A4 size. A flexible Reader using E Ink, which is
planned to reach the market early in 2010, is made by Plastic Logic in
Dresden. This incorporates polymer TFTs (Burns et al. 2006).
Production
of a rival EP display was announced by Bridgestone in April 2009. The
medium is a ‘liquid powder’, and this is held between two plastic
plates, with barriers between the plates to prevent pigment motion (Sakurai et al. 2006). Small colour displays have been used for supermarket labels, and now displays of A4 size are coming on to the market.
Though
constraining the volume available to particle diffusion in EP displays
can certainly reduce clumping and gravitational settling, it does not
eliminate them, and electrode staining, with a higher voltage needed to
remove particles than in first use, is still a problem. There is little
published on the life of electronic paper based on EP, but the patent
literature shows that the reflectivity of early models halved within
2000 h. Modern paper will show some improvement on that figure.
One
requirement of visual effects for electronic paper is bistability,
since power is then required only when the display needs updating. If
the bistable states are separated by a significant energy barrier, so
that transition requires a short pulse with voltage above a threshold,
passive matrix addressing is possible, with much simplification in back
plane design, and a considerable reduction in processing cost. Though
standard TN LCDs are not bistable, a change in the design can give this
property. ZBD Displays, following earlier research on LC grating
alignment (Flanders et al. 1978), adds a grating structure to one surface of a TN display, and there are then two stable states (figure 34).
Switching from one state to the other requires a short pulse, so the
updating of a small page requires only 250 ms. The display has a
reflectivity of 39 per cent and a contrast ratio of 10 : 1, surprisingly
good figures considering the need for polarizers (Jones 2007). The main use of the technology is for shelf-edge labelling (figure 35), but there seems no technical obstacle to an increase in size.
Bistable LC states in ZBD display.
Nemoptics
obtains a similar effect without the grating. They treat one surface so
that the anchoring of the LC molecules is weak, and obtain two stable
states with the nematic molecules parallel to the surfaces, one twisted
and one aligned. Molecules can move from one state to the other only by
passing through a third state, with molecules orthogonal to the
surfaces. This transition is made by unipolar pulses of different shapes
for writing and erasing (Faget et al. 2006; Dozov et al. 2007).
The display reflectivity is 33 per cent, with a contrast ratio of over
10 : 1. An A4 page can be refreshed on 0.4 s. Nemoptics displays are
available in colour, under the trade name BiNem, and are being used for
shelf labelling. In November 2007 they announced that Seiko Instruments
would be using BiNem for a Reader.
Though both ZBD and
Nemoptics have made experimental displays on plastic substrates, there
is no information on life. The glass versions have the life of normal
LCDs, and, where weight and flexibility are not a prerequisite, could
provide an E-book with a simpler back plane, presumably at a lower cost
than the electrophoretic versions.
This
is a dynamic field, and while EP has produced commercial devices, and
bistable TN has demonstrated small working prototypes, there are many
reports of visual effects that are claimed to have superior properties
and immense potential for E-books. Proposed effects include
electrowetting, electrofluidics and bistability in cholesteric LCs. Many
mechanical effects in microelectromechanical systems (MEMS) are also
under consideration.
(d) Other applications
There
are a number of additional applications of flat-panel electronic
displays, but though the uses differ greatly from those covered earlier,
the physics and chemistry are similar. The exception is the very large
displays, such as those used for video presentation in sports stadia.
They use several million LEDs, generally in tiles about a metre square.
Smaller displays used for signs and billboards can exploit cholesteric
LCDs. The sign shown in figure 36 is constructed from over 200 tiles to give an area of over 10 m2, with a reflectivity of 32 per cent. The picture can change in 1–2 s (Coates 2008).
Any
flat-panel display lacks reality unless it can show three dimensions.
The simplest method requires the user to wear polarized spectacles, a
disadvantage for many users. Most other attempts give a good
stereoscopic image to one observer, but others can see a disturbing
pseudoscopic presentation. The most convincing representation comes from
computer-generated holography, but this involves much computation, and
is very expensive.
7. Future trends
Market
predictions are always suspect, because they assume that consumer
behaviour is like ‘the law of the Medes and Persians, which altereth
not’ (King James Bible, Daniel 6:8). Predictions based on
technological progress are even more questionable. Nevertheless, a
survey of flat-panel history and achievement would be incomplete without
some words on the future. Some developments seem clear. The TV market
will see competition between LCDs and the various forms of OLED, with
plasma panels decreasing in importance. The conventional CRT will die
away, and it is unlikely that any flat form will become significant.
Competition
in smaller devices is more difficult to assess. Flexibility can be an
important feature in design, but its advantages pale if linked with a
short display life. Nevertheless, if a fashionable item is inexpensive,
its life is peripheral. That would indicate an opening for flexible
OLEDs in mobile phones, but problems in Readers. The limited success to
date of the research on barrier layers gives concern, because much
effort and funds have been applied. If the reports of success from
Singapore remain unconfirmed, glass Readers would prevail, with a
possible opening for passive matrix bistable LCDs.
Silicon
is truly a workhorse in electronics, and its versatility has been shown
in the Texas Instruments Digital Light Projector. The technology has a
natural home in microdisplays, and it is certainly possible that
interesting optical effects can be obtained by sub-micrometre motion of
layers in MEMS. Ingenuity will be needed, however, before the high
definition possible in small chips can be projected to larger areas
within the limited depth of Readers.
A future without
three-dimensional TV displays is unimaginable, and a few companies have
made experimental models that come close to a satisfactory presentation.
The interim steps of the Dolby three-dimensional Digital Cinema, using
polarizing goggles, and the Planar Stereo Monitor, involving two
screens, may well kindle consumer interest enough for companies to
market displays that approximate to reality in a form acceptable to
consumers.
The past 40 years have
seen immense progress in flat-panel technology, and has led to huge
funds being applied in the many areas where panels can provide excellent
images. The future is likely to be as bright for displays as the past
has been rewarding for physicists, chemists and engineers.
High-fidelity electronic display of digital radiographs.
A fully digital radiography system requires high-fidelity electronic
display devices that preserve diagnostic quality. Current cathode-ray
tube monitors do not meet desired performance criteria for displaying
radiographs and have excessive size, weight, and power consumption.
Recent developments in flat-panel display technology (in particular
active-matrix liquid crystal displays, field-emission displays, and
organic light-emitting displays) suggest that high-fidelity, lightweight
displays will be available in the near future. Large-size active-matrix
liquid crystal display devices have been demonstrated. High brightness
can be easily achieved with bright back illumination. Further
developments in optical design for monochrome displays should provide
high fidelity and improve the angular dependence of the emitted light.
Field-emission display devices have attractive emission distribution and
potential for low veiling glare. This technology needs to be extended
to a large area, and problems with cathode aging and nonuniformity have
to be contemplated. Organic light-emitting displays represent a simple
and potentially inexpensive display technology with the ability to
achieve high image quality. However, extensive research and development
is required to achieve large-area manufacturing methods.
Concept AUTOMATION
Monitoring and reporting ensure the opportunity to access all control points from any given point in the system as well as flexibility, speed and minimized faults. Kontrolmatik maximizes automation through ”Balance of Plant“ integration, control systems and turbine control systems specifically designed for your facility.
Contributions of automation systems to power plants;
- Opportunity to access all control parameters from any point in the system
- Managing and monitoring the entire system from a single control platform
- Rapid command issuing
- Integration of electric infrastructure and automation systems
- Prior determination and prevention of unexpected faults and reduction of maintenance time
- Secure data acquisition and fast data flow
Light Sensors
Light Sensors are photoelectric devices that convert light energy (photons) whether visible or infra-red light into an electrical (electrons) signal .A Light Sensor generates an output signal indicating the intensity of light by measuring the radiant energy that exists in a very narrow range of frequencies basically called “light”, and which ranges in frequency from “Infra-red” to “Visible” up to “Ultraviolet” light spectrum.
The light sensor is a passive devices that convert this “light energy” whether visible or in the infra-red parts of the spectrum into an electrical signal output. Light sensors are more commonly known as “Photoelectric Devices” or “Photo Sensors” because the convert light energy (photons) into electricity (electrons).
Photoelectric devices can be grouped into two main categories, those which generate electricity when illuminated, such as Photo-voltaics or Photo-emissives etc, and those which change their electrical properties in some way such as Photo-resistors or Photo-conductors. This leads to the following classification of devices.
- • Photo-emissive Cells – These are photodevices which release free electrons from a light sensitive material such as caesium when struck by a photon of sufficient energy. The amount of energy the photons have depends on the frequency of the light and the higher the frequency, the more energy the photons have converting light energy into electrical energy.
- • Photo-conductive Cells – These photodevices vary their electrical resistance when subjected to light. Photoconductivity results from light hitting a semiconductor material which controls the current flow through it. Thus, more light increase the current for a given applied voltage. The most common photoconductive material is Cadmium Sulphide used in LDR photocells.
- • Photo-voltaic Cells – These photodevices generate an emf in proportion to the radiant light energy received and is similar in effect to photoconductivity. Light energy falls on to two semiconductor materials sandwiched together creating a voltage of approximately 0.5V. The most common photovoltaic material is Selenium used in solar cells.
- • Photo-junction Devices – These photodevices are mainly true semiconductor devices such as the photodiode or phototransistor which use light to control the flow of electrons and holes across their PN-junction. Photojunction devices are specifically designed for detector application and light penetration with their spectral response tuned to the wavelength of incident light.
The Photoconductive Cell
A Photoconductive light sensor does not produce electricity but simply changes its physical properties when subjected to light energy. The most common type of photoconductive device is the Photoresistor which changes its electrical resistance in response to changes in the light intensity.Photoresistors are Semiconductor devices that use light energy to control the flow of electrons, and hence the current flowing through them. The commonly used Photoconductive Cell is called the Light Dependent Resistor or LDR.
The Light Dependent Resistor
Typical LDRThe net effect is an improvement in its conductivity with a decrease in resistance for an increase in illumination. Also, photoresistive cells have a long response time requiring many seconds to respond to a change in the light intensity.
Materials used as the semiconductor substrate include, lead sulphide (PbS), lead selenide (PbSe), indium antimonide (InSb) which detect light in the infra-red range with the most commonly used of all photoresistive light sensors being Cadmium Sulphide (Cds).
Cadmium sulphide is used in the manufacture of photoconductive cells because its spectral response curve closely matches that of the human eye and can even be controlled using a simple torch as a light source. Typically then, it has a peak sensitivity wavelength (λp) of about 560nm to 600nm in the visible spectral range.
The Light Dependent Resistor Cell
To increase the dark resistance and therefore reduce the dark current, the resistive path forms a zigzag pattern across the ceramic substrate. The CdS photocell is a very low cost device often used in auto dimming, darkness or twilight detection for turning the street lights “ON” and “OFF”, and for photographic exposure meter type applications.
The amount of voltage drop across series resistor, R2 is determined by the resistive value of the light dependant resistor, RLDR. This ability to generate different voltages produces a very handy circuit called a “Potential Divider” or Voltage Divider Network.
As we know, the current through a series circuit is common and as the LDR changes its resistive value due to the light intensity, the voltage present at VOUT will be determined by the voltage divider formula. An LDR’s resistance, RLDR can vary from about 100Ω in the sun light, to over 10MΩ in absolute darkness with this variation of resistance being converted into a voltage variation at VOUT as shown.
One simple use of a Light Dependent Resistor, is as a light sensitive switch as shown below.
LDR SwitchAs the light level increases the resistance of the LDR starts to decrease causing the base bias voltage at V1 to rise. At some point determined by the potential divider network formed with resistor R1, the base bias voltage is high enough to turn the transistor TR1 “ON” and thus activate the relay which in turn is used to control some external circuitry. As the light level falls back to darkness again the resistance of the LDR increases causing the base voltage of the transistor to decrease, turning the transistor and relay “OFF” at a fixed light level determined again by the potential divider network.
By replacing the fixed resistor R1 with a potentiometer VR1, the point at which the relay turns “ON” or “OFF” can be pre-set to a particular light level. This type of simple circuit shown above has a fairly low sensitivity and its switching point may not be consistent due to variations in either temperature or the supply voltage. A more sensitive precision light activated circuit can be easily made by incorporating the LDR into a “Wheatstone Bridge” arrangement and replacing the transistor with an Operational Amplifier as shown.
Light Level Sensing Circuit
The operational amplifier is configured as a Differential Amplifier also known as a voltage comparator with feedback whose output voltage condition is determined by the difference between the two input signals or voltages, V1 and V2. The resistor combination R1 and R2 form a fixed voltage reference at input V2, set by the ratio of the two resistors. The LDR – VR1 combination provides a variable voltage input V1 proportional to the light level being detected by the photoresistor.
As with the previous circuit the output from the operational amplifier is used to control a relay, which is protected by a free wheel diode, D1. When the light level sensed by the LDR and its output voltage falls below the reference voltage set at V2 the output from the op-amp changes state activating the relay and switching the connected load.
Likewise as the light level increases the output will switch back turning “OFF” the relay. The hysteresis of the two switching points is set by the feedback resistor Rf can be chosen to give any suitable voltage gain of the amplifier.
The operation of this type of light sensor circuit can also be reversed to switch the relay “ON” when the light level exceeds the reference voltage level and vice versa by reversing the positions of the light sensor LDR and the potentiometer VR1. The potentiometer can be used to “pre-set” the switching point of the differential amplifier to any particular light level making it ideal as a simple light sensor project circuit.
Photojunction Devices
Photojunction Devices are basically PN-Junction light sensors or detectors made from silicon semiconductor PN-junctions which are sensitive to light and which can detect both visible light and infra-red light levels. Photo-junction devices are specifically made for sensing light and this class of photoelectric light sensors include the Photodiode and the Phototransistor.The Photodiode.
Photo-diodeThis characteristic can be a problem for diodes with transparent or glass bead bodies such as the 1N4148 signal diode. LED’s can also be used as photodiodes as they can both emit and detect light from their junction. All PN-junctions are light sensitive and can be used in a photo-conductive unbiased voltage mode with the PN-junction of the photodiode always “Reverse Biased” so that only the diodes leakage or dark current can flow.
The current-voltage characteristic (I/V Curves) of a photodiode with no light on its junction (dark mode) is very similar to a normal signal or rectifying diode. When the photodiode is forward biased, there is an exponential increase in the current, the same as for a normal diode. When a reverse bias is applied, a small reverse saturation current appears which causes an increase of the depletion region, which is the sensitive part of the junction. Photodiodes can also be connected in a current mode using a fixed bias voltage across the junction. The current mode is very linear over a wide range.
Photo-diode Construction and Characteristics
Thus, the photodiodes current is directly proportional to light intensity falling onto the PN-junction. One main advantage of photodiodes when used as light sensors is their fast response to changes in the light levels, but one disadvantage of this type of photodevice is the relatively small current flow even when fully lit.
The following circuit shows a photo-current-to-voltage converter circuit using an operational amplifier as the amplifying device. The output voltage (Vout) is given as Vout = IP*Rƒ and which is proportional to the light intensity characteristics of the photodiode.
This type of circuit also utilizes the characteristics of an operational amplifier with two input terminals at about zero voltage to operate the photodiode without bias. This zero-bias op-amp configuration gives a high impedance loading to the photodiode resulting in less influence by dark current and a wider linear range of the photocurrent relative to the radiant light intensity. Capacitor Cf is used to prevent oscillation or gain peaking and to set the output bandwidth (1/2πRC).
Photo-diode Amplifier Circuit
The Phototransistor
Photo-transistorPhototransistors operate the same as the photodiode except that they can provide current gain and are much more sensitive than the photodiode with currents are 50 to 100 times greater than that of the standard photodiode and any normal transistor can be easily converted into a phototransistor light sensor by connecting a photodiode between the collector and base.
Phototransistors consist mainly of a bipolar NPN Transistor with its large base region electrically unconnected, although some phototransistors allow a base connection to control the sensitivity, and which uses photons of light to generate a base current which in turn causes a collector to emitter current to flow. Most phototransistors are NPN types whose outer casing is either transparent or has a clear lens to focus the light onto the base junction for increased sensitivity.
Photo-transistor Construction and Characteristics
Usually the sensitivity of a phototransistor is a function of the DC current gain of the transistor. Therefore, the overall sensitivity is a function of collector current and can be controlled by connecting a resistance between the base and the emitter but for very high sensitivity optocoupler type applications, Darlington phototransistors are generally used.
Photo-darlingtonPhoto darlington devices consist of a normal phototransistor whose emitter output is coupled to the base of a larger bipolar NPN transistor. Because a darlington transistor configuration gives a current gain equal to a product of the current gains of two individual transistors, a photodarlington device produces a very sensitive detector.
Typical applications of Phototransistors light sensors are in opto-isolators, slotted opto switches, light beam sensors, fibre optics and TV type remote controls, etc. Infrared filters are sometimes required when detecting visible light.
Another type of photojunction semiconductor light sensor worth a mention is the Photo-thyristor. This is a light activated thyristor or Silicon Controlled Rectifier, SCR that can be used as a light activated switch in AC applications. However their sensitivity is usually very low compared to equivalent photodiodes or phototransistors.
To help increase their sensitivity to light, photo-thyristors are made thinner around the gate junction. The downside to this process is that it limits the amount of anode current that they can switch. Then for higher current AC applications they are used as pilot devices in opto-couplers to switch larger more conventional thyristors.
Photovoltaic Cells.
The most common type of photovoltaic light sensor is the Solar Cell. Solar cells convert light energy directly into DC electrical energy in the form of a voltage or current to a power a resistive load such as a light, battery or motor. Then photovoltaic cells are similar in many ways to a battery because they supply DC power.However, unlike the other photo devices we have looked at above which use light intensity even from a torch to operate, photovoltaic solar cells work best using the suns radiant energy.
Solar cells are used in many different types of applications to offer an alternative power source from conventional batteries, such as in calculators, satellites and now in homes offering a form of renewable power.
Photovoltaic CellWhen illuminated the light energy causes electrons to flow through the PN junction and an individual solar cell can generate an open circuit voltage of about 0.58v (580mV). Solar cells have a “Positive” and a “Negative” side just like a battery.
Individual solar cells can be connected together in series to form solar panels which increases the output voltage or connected together in parallel to increase the available current. Commercially available solar panels are rated in Watts, which is the product of the output voltage and current (Volts times Amps) when fully lit.
Characteristics of a typical Photovoltaic Solar Cell.
Other materials used in the construction of photovoltaic cells include Gallium Arsenide, Copper Indium Diselenide and Cadmium Telluride. These different materials each have a different spectrum band response, and so can be “tuned” to produce an output voltage at different wavelengths of light.
In this tutorial about Light Sensors, we have looked at several examples of devices that are classed as Light Sensors. This includes those with and those without PN-junctions that can be used to measure the intensity of light.
In the next tutorial we will look at output devices called Actuators. Actuators convert an electrical signal into a corresponding physical quantity such as movement, force, or sound. One such commonly used output device is the Electromagnetic Relay.
Sound Transducers
Sound Transducers use electrical energy to create mechanical vibrations to disturbe the sourrounding air producing sound whether of an audible or inaudible frequency .Sound is the generalised name given to “acoustic waves”. These acoustic waves have frequencies ranging from just 1Hz up to many tens of thousands of Hertz with the upper limit of human hearing being around the 20 kHz, (20,000Hz) range.
The sound that we hear is basically made up from mechanical vibrations produced by an Audio Sound Transducer used to generate the acoustic waves, and for sound to be “heard” it requires a medium for transmission either through the air, a liquid, or a solid.
Sound TransducerAudio Sound Transducers include both input sensors, that convert sound into and electrical signal such as a microphone, and output actuators that convert the electrical signals back into sound such as a loudspeaker.
We tend to think of sound as only existing in the range of frequencies detectable by the human ear, from 20Hz up to 20kHz (a typical loudspeaker frequency response), but sound can also extend way beyond these ranges.
Sound transducers can also both detect and transmit sound waves and vibrations from very low frequencies called infra-sound up to very high frequencies called ultrasound. But in order for a sound transducer to either detect or produce “sound” we first need to understand what sound is.
What is Sound?
Sound is basically a waveform of energy that is produced by some form of a mechanical vibration such as a tuning fork, and which has a “frequency” determined by the origin of the sound for example, a bass drum has a low frequency sound while a cymbal has a higher frequency sound.A sound waveform has the same characteristics as that of an electrical waveform which are Wavelength (λ), Frequency (ƒ) and Velocity (m/s). Both the sounds frequency and wave shape are determined by the origin or vibration that originally produced the sound but the velocity is dependent upon the medium of transmission (air, water etc.) that carries the sound wave. The relationship between wavelength, velocity and frequency is given below as:
Sound Wave Relationship
- Where:
- Wavelength – is the time period of one complete cycle in Seconds, (λ)
- Frequency – is the number of wavelengths per second in Hertz, (ƒ)
- Velocity – is the speed of sound through a transmission medium in m/s-1
The Microphone Input Transducer
The Microphone, also called a “mic”, is a sound transducer that can be classed as a “sound sensor”. This is because it produces an electrical analogue output signal which is proportional to the “acoustic” sound wave acting upon its flexible diaphragm. This signal is an “electrical image” representing the characteristics of the acoustic waveform. Generally, the output signal from a microphone is an analogue signal either in the form of a voltage or current which is proportional to the actual sound wave.The most common types of microphones available as sound transducers are Dynamic, Electret Condenser, Ribbon and the newer Piezo-electric Crystal types. Typical applications for microphones as a sound transducer include audio recording, reproduction, broadcasting as well as telephones, television, digital computer recording and body scanners, where ultrasound is used in medical applications. An example of a simple “Dynamic” microphone is shown below.
Dynamic Moving-coil Microphone Sound Transducer
The movement of the coil within the magnetic field causes a voltage to be induced in the coil as defined by Faraday’s law of Electromagnetic Induction. The resultant output voltage signal from the coil is proportional to the pressure of the sound wave acting upon the diaphragm so the louder or stronger the sound wave the larger the output signal will be, making this type of microphone design pressure sensitive.
As the coil of wire is usually very small the range of movement of the coil and attached diaphragm is also very small producing a very linear output signal which is 90o out of phase to the sound signal. Also, because the coil is a low impedance inductor, the output voltage signal is also very low so some form of “pre-amplification” of the signal is required.
As the construction of this type of microphone resembles that of a loudspeaker, it is also possible to use an actual loudspeaker as a microphone.
Obviously, the average quality of a loudspeaker will not be as good as that for a studio type recording microphone but the frequency response of a reasonable speaker is actually better than that of a cheap “freebie” microphone. Also the coils impedance of a typical loudspeaker is different at between 8 to 16Ω. Common applications where speakers are generally used as microphones are in intercoms and walki-talkie’s.
The Loudspeaker Output Transducer
Sound can also be used as an output device to produce an alert noise or act as an alarm, and loudspeakers, buzzers, horns and sounders are all types of sound transducer that can be used for this purpose with the most commonly used audible type output sound actuator being the “Loudspeaker”.
Loudspeaker TransducerLoudspeakers are available in all shapes, sizes and frequency ranges with the more common types being moving coil, electrostatic, isodynamic and piezo-electric. Moving coil type loudspeakers are by far the most commonly used speaker in electronic circuits, kits and toys, and as such it is this type of sound transducer we will examine below.
The principle of operation of the Moving Coil Loudspeaker is the exact opposite to that of the “Dynamic Microphone” we look at above. A coil of fine wire, called the “speech or voice coil”, is suspended within a very strong magnetic field, and is attached to a paper or Mylar cone, called a “diaphragm” which itself is suspended at its edges to a metal frame or chassis. Then unlike the microphone which is pressure sensitive input device, this type of sound transducer can be classed as a pressure generating output device.
The Moving Coil Loudspeaker
As the voice coil is permanently attached to the cone/diaphragm this also moves in tandem and its movement causes a disturbance in the air around it thus producing a sound or note. If the input signal is a continuous sine wave then the cone will move in and out acting like a piston pushing and pulling the air as it moves and a continuous single tone will be heard representing the frequency of the signal. The strength and therefore its velocity, by which the cone moves and pushes the surrounding air produces the loudness of the sound.
As the speech or voice coil is essentially a coil of wire it has, like an inductor an impedance value. This value for most loudspeakers is between 4 and 16Ω and is called the “nominal impedance” value of the speaker measured at 0Hz, or DC.
Remember that it is important to always match the output impedance of the amplifier with the nominal impedance of the speaker to obtain maximum power transfer between the amplifier and speaker. Most amplifier-speaker combinations have an efficiency rating as low as 1 or 2%.
Although disputed by some, the selection of good speaker cable is also an important factor in the efficiency of the speaker, as the internal capacitance and magnetic flux characteristics of the cable change with the signal frequency, thereby causing both frequency and phase distortion. This has the effect of attenuating the signal. Also, with high power amplifiers large currents are flowing through these cables so small thin bell wire type cables can overheat during extended periods of use, again reducing efficiency.
The human ear can generally hear sounds from between 20Hz to 20kHz, and the frequency response of modern loudspeakers called general purpose speakers are tailored to operate within this frequency range as well as headphones, earphones and other types of commercially available headsets used as sound transducers.
However, for high performance High Fidelity (Hi-Fi) type audio systems, the frequency response of the sound is split up into different smaller sub-frequencies thereby improving both the loudspeakers efficiency and overall sound quality as follows:
Generalised Frequency Ranges
| Descriptive Unit | Frequency Range |
| Sub-Woofer | 10Hz to 100Hz |
| Bass | 20Hz to 3kHz |
| Mid-Range | 1kHz to 10kHz |
| Tweeter | 3kHz to 30kHz |
This crossover network consists of Resistors, Inductors, Capacitors, RLC type passive filters or op-amp active filters whose crossover or cut-off frequency point is finely tuned to that of the individual loudspeakers characteristics and an example of a multi-speaker “Hi-fi” type design is given below.
Multi-speaker (Hi-Fi) Design
XO__XO CPU amplifier
Some
of the better — albeit, not the greatest, but definitely better —
solutions I’ve seen for speeding up a laptop when it’s docked are those
external water-cooling blocks. This will directly improve the
performance of your native CPU by providing better cooling to the existing CPU.
The
reason you can have such a great benefit from an external GPU (eGPU)
and not as much of one with an external “CPU” is because the graphics
card is doing all of its own work in-house, and shipping the finished
product back. A CPU on the other hand is in constant communication with
various components, and needs to have the snappiest connection possible.
It
doesn’t necessarily need a big ol’ highway with lots of traffic on
which a postal truck can carry packages. Instead, it needs a telephone
or radio that instantly connects, sends a little info, gets an answer,
does a little work, and repeats this.
The graphics card is the big honkin’ highway that gets a massive request for a ton of work to be done in one burst, and then sits down and number crunches on its own for a while. It ships out product whenever it can, regardless of the receiving unit’s readiness. Extra product is ignored, and shortages are met with stuttering displays and the occasional rage-induced punch to the screen.
The graphics card is the big honkin’ highway that gets a massive request for a ton of work to be done in one burst, and then sits down and number crunches on its own for a while. It ships out product whenever it can, regardless of the receiving unit’s readiness. Extra product is ignored, and shortages are met with stuttering displays and the occasional rage-induced punch to the screen.
The
only way to get the snappy connection is to have the unit physically as
close together as possible. The big highway analogy can be accomplished
with a single high-bandwidth cable.
A compute node is great for tasks that can be shipped out and received “whenever.” Most of your everyday tasks are not this.
the CPU is central to the system and physically must be close as
possible to the motherboard. Operations are occurring so fast than the
physical length of the traces/wires matter.
There
are co-processors that plug into the motherboard and off-load specific
workloads from the CPU…. video cards, sound cards, enterprise storage
controllers, and enterprise network cards are examples. They are
designed to use an interconnect which shuttles data to and from the CPU
in packets. Note that the GPU core is attached to a circuit board itself
with its own memory and power system. These expansion cards all work
the same way…. a bunch of data gets shipped to the card, the card
processes the data onboard itself, then ships the batch of data out.
You
can sometimes use “different” CPUs in a system. Occassionaly, AMD or
Intel physically locks out a CPU from an older system with an extra pin.
However, this has happened only a handful of times and it really isn’t a
different CPU… just a different class of the same design.
If
you want to increase the performance of your CPU, you might be able to
overclock it. This can be done by increasing the base clock on your
system or increasing the clock multiplier to increase the CPUs clock in
relation to the base clock.
Power Request
Voltage converter
A voltage converter is an electric power converter which changes the voltage of an electrical power source. It may be combined with other components to create a power supply.
AC and DC
AC voltage conversion uses a transformer. Conversion from one DC voltage to another requires electronic circuitry(electromechanical equipment was required before the development of semiconductor electronics), like a DC-DC converter. Mains power (called household current in the US) is universally AC.Practical voltage converters
Mains converters
A common use of the voltage converter is for a device that allows appliances made for the mains voltage of one geographical region to operate in an area with different voltage. Such a device may be called a voltage converter, power converter, travel adapter, etc. Most single phase alternating-current electrical outlets in the world supply power at 210–240 V or at 100–120 V. A transformer or autotransformer can be used; (auto)transformers are inherently reversible, so the same transformer can be used to step the voltage up, or step it down by the same ratio. Lighter and smaller devices can be made using electronic circuitry; reducing the voltage electronically is simpler and cheaper than increasing it. Small, inexpensive, travel adapters suitable for low-power devices such as electric shavers, but not, say, hairdriers, are available; travel adapters usually include plug-end adapters for the different standards used in different countries. A transformer would be used for higher power.Transformers do not change the frequency of electricity; in many regions with 100–120 V, electricity is supplied at 60 Hz, and 210–240 V regions tend to use 50 Hz. This may affect operation of devices which depend on mains frequency (some audio turntables and mains-only electric clocks, etc., although modern equipment is less likely to depend upon mains frequency). Equipment with high-powered motors or internal transformers designed to operate at 60 Hz may overheat at 50 Hz even if the voltage supplied is correct.
Most mains-powered electrical equipment, though it may specify a single nominal voltage, actually has a range of tolerance above and below that point. Thus, devices usually can be used on either any voltage from approx. 100 to 120 V, or any voltage from approx. 210 to 240 V. In such cases, voltage converters need only be specified to convert any voltage within one range, to a voltage within the other, rather than separate converters being needed for all possible pairs of nominal voltages (110–220, 117–220, 110–230, etc.)
Converters for devices
Mains converters
Another requirement is to provide low-voltage electricity to a device from mains electricity; this would be done by what is usually called a power supply. Most modern electronic devices require between 1.5 and 24 volts DC; lower-powered devices at these voltages can often work either from batteries or mains. Some devices incorporate a power supply and are simply plugged into the mains. Others use an external power supply comprising either a transformer and rectifier, or electronic circuitry. Switched-mode power supplies have become widespread in the early twenty-first century; they are smaller and lighter than the once-universal transformer converters, and are often designed to work from AC mains at any voltage between 100 and 250 V. Additionally, because they are typically rectified to operate at a DC voltage, they are minimally affected by the frequency of the mains (50 vs 60 Hz). Details on operation are given in the article on power supplies.Mobile converters
Voltage converters can be used in vehicles with 12 V DC outlets. A simple voltage dropper can be used to reduce the voltage for low-power devices; if more than 12V is required, or for high-powered devices, a switched-mode power supply is used. The output will usually be DC in the range 1.5–24 V. Power supplies that output either 100–120 V AC or 210–240 V AC are available; they are called inverters, due to the conversion from DC to AC rather than the voltage change. The output frequency and waveform of an inverter may not accurately replicate that supplied by mains electricity, although this is not usually a problem.Choosing a converter
A converter to drive equipment should be specified to supply at least the actual watts or amperes used by the equipment; this is usually stated on a label on the equipment. A converter of higher rating is safe to use. Some equipment uses more power or current at startup; a 20% margin is usually adequate, although it may be included in the rated figure.MACRO SYSTEM
|
MACRO is an acronym for "Motion and Control Ring Optical". This is a non-proprietary digital interface developed by Delta Tau Data Systems for connection of multi-axis motion controllers, amplifiers, and I/O. MACRO is an Ethernet style protocol that uses a multi-mode fiber optic (FDDI-PMD ISO/IEC 9314-3) or twisted pair copper (CAT5) which is called 100BASEFX and 100BASETX. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Single-plug connections between controls and amps minimize the wiring complexity in large systems. A single fiber optic strand can provide a controller with: position feedback, flag status (limits, home flag, registration prox status), amp status, and machine input status. This same strand can communicate to the amplifier and other devices on the MACRO network. (Amp enable & amp command signals, machine outputs, commands to D/A converters, all can be implemented with a single plug connection.) Setup and initialization information for amps and other devices also can be transmitted through the strand, eliminating in many cases the need for separate RS232 cables. Less wiring translates into faster machine assembly, lower assembly costs, fewer errors, and easier trouble-shooting in the field. Noise Immunity. Fiber-optic cable transmits light, not electricity. Unlike electricity, light is immune to electromagnetic noise, capacitive coupling, ground loops, and other wiring nightmares. The copper based system is transformer isolated for better noise immunity. Speed. One of MACRO's most impressive features is it's 125 Mbits/Sec rate of data transmission. This is at least 25 times faster than most other digital motion control interfaces. What good is all this speed? Well, take the case of a system with 16 amplifiers and two 8-axis controllers. Updates to each amplifier and controller can occur at less than 25 m sec intervals (>40kHz). This is more than fast enough to close high performance servo loops across the MACRO ring, allowing the flexibility to choose distributed intelligence or centralized control! Digital communications allows the ability to "tie it all together". In a motion control application there are a number of different tasks: trajectory planning, position loop closure, velocity loop closure, current loop closure, and commutation (if the motor is Brushless or AC Induction - flux-vector controlled). Controllers have always performed trajectory planning and position loop closure digitally. Today, many controllers are also closing the velocity loops digitally as well. Amplifiers are also changing with the times. AC Servo and Vector amplifiers have always performed their commutation functions digitally; however several amplifiers now close their current loops digitally as well. Digital communications allows the ability to tie all these tasks together without conversion to the analog realm. This can improve system performance by minimizing error in D-to-A, and A-to-D conversion, eliminating ground loops and voltage offset/shift problems. It also allows adaptive control techniques, enhances the ability to analyze & critique system performance using techniques like data-gathering, and even provides the potential to auto-diagnose a system problem from a central or supervisory control. One ring, multiple masters. In a ring topology, a central control serially transmits information to several nodes (see illustration). Today's multi-axis controls typically support only 4 to 8 axes. In machines with more than 8 axes, having a ring for each control is sometimes not desirable. In these cases MACRO will support the ability to put several of these controllers ("masters") together on one ring, further simplifying the system complexity. Simplicity. Transmission across the MACRO ring requires no software intervention. The information to be sent from one node to another is merely written to a memory location, and the MACRO hardware takes care of the rest. A programmer does not need to worry about slipping "telegrams" into the timing of the ring communications, or interrupting a task to update the ring.
What
industries/applications does MACRO support?
MACRO lends itself well to large multi-axis applications where the amplifiers and I/O are spread out in several cabinets around the machine. Packaging, converting, processing, and textiles machinery as well as robotics systems, and automated assembly lines all can benefit from MACRO's advantages. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
MACRO lends itself well to large multi-axis applications where the amplifiers and I/O are spread out in several cabinets around the machine. Packaging, converting, processing, and textiles machinery as well as robotics systems, and automated assembly lines all can benefit from MACRO's advantages. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Schematically, what does a MACRO network look like? MACRO operates in a ring topology (see above sketch). Data is transmitted serially. Each station on the ring has an "in" port for receiving data and an "out" port for transmitting data. Nodes, residing at a station, can be amplifier axes, I/O banks, or communication interfaces to other devices. A station can have one or several nodes, allowing for multi-axis amplifiers with a single "in" and single "out" port. Data packets, (groups of 96 bits of serial data), sent out by a motion controller or, master node, are earmarked for a specific amplifier, or slave node. If the data packet is not for an amplifier, it passes it on, unchanged. If it is for the node, it copies the contents of the data packet (typically commands), places feedback data into the packet in place of the command data, and transmits the data packet. What happens at a node? The operation of MACRO is based on standard FDDI (Fiber Direct Digital Interface) networking technology. In the illustration below, data written to MACRO from a processor, is shifted out via byte wide shift registers. These shift registers (typically embedded in an ASIC) send bytes of data to a Network Interface Driver IC which serializes the bytes of data, and outputs a serial stream at 125 Mbits/sec. The network driver adds 2 bits of error detection to the byte and puts out the data in a 4B/5B coding scheme. This serialized digital data is then routed through an optical or electrical driver for transmission to remote nodes. An address specifier at the beginning of data packets indicates whether or not the data following in the packet is intended for the node. If it is not, it is put right back onto the ring via the process just described and graphically illustrated as Case 1 below. If however, the data is for the node, a simultaneous "copy and replace" process is performed as shown in Case 2. Since the MACRO shift registers manage the communications (shifting data out and in to the ring at the appropriate times), the network transmission is transparent to the controller software. If information is required, a register is simply read. If there is a need to transmit information, it is merely written to a register... MACRO hardware does the rest! This parallel register interaction makes MACRO quick, simple and efficient for a processor to interact with the MACRO ring. |
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
As with any type of communications, both the transmitting device and the receiving device must "speak the same language". However, MACRO has the ability to support different protocols for different devices. For example, a controller can communicate to several amplifiers, each of which expects a different type of command. The only requirement is that the controller and amp be preset to respectively deliver and accept the same type of data. (This too can be done over the ring.) The charts below illustrate the TYPE 1 protocol for different types of amplifier interfaces and I/O equipment interfaces. Register 0 is a 24-bit (3-byte) register. Registers 1, 2, and 3 are 16-bit (2-byte) registers.
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
e- Combining and connecting REWIND and signals
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Tidak ada komentar:
Posting Komentar