- Email: [email protected]
Instrument technology of the future
Instrument technology of the future
of taxpayers' money to encourage their manufacture and application.
R e g M e d l o c k o f K e n t Instruments examines p o s s i b l e areas o f advancement f o r i n s t r u m e n t technology, particularly in improving t h e e f f i c i e n c y a n d s a f e t y o f t h e process industries
Microelectronic
The process industries have a tradition of conservatism towards the adoption of new products whose performance, reliability and back-up resources are still to be proved. Furthermore, through replacement, retrofit and user familiarity, the life of older products is extended, sometimes to Obsolescence, thereby retarding the acceptance of newer developments. The effect of this is that the prophets of technology tend to overestimate progress in the first five years but underestimate in the following 20 years. M ICROE L ECTRON
ICS
There is however, a unanimity of opinion amongst forecasters that the instrument industry will be dominated by microelectronics for the foreseeable future. In this article microelectronics covers microprocessors, memories, silicon chip devices, integrated circuits, LSI and VLSI. The news media frequently refer to microprocessors when they really mean microcomputers or even the broader field of microelectronics. The precise boundary between microprocessors and microcomputers is difficult to establish as it dep'ends on what has been included on the microprocessor chip. Normally a microprocessor is a component and a microcomputer is an assembled kit on a printed circuit board comprising one or more microprocessor units and additional components such as memories. The growth of computers in industrial control was slow I~o start but has begun to reach explosive proportions. It has been estimated that, since the first computer control applications in 1960, half a million central processors have been installed (not just on process control of course). The growth rato has averaged 67% per annum and it is believed that 240 000 processors
474
applications
Microcircuitry including microprocessors and other microcircuit chips will be applied to instruments in many ways. Table 1 shows a n u ~ of.e~amples. In the excitement of microprocessor were added in ] 978 and 340 000 will application it is sometimes forgotten be added in 1979. that usage will be limited by problems Microcomputers look as if they associated with interfacing, software will invade all industry as well as the production and mechanisms in the domestic market. Retail shops in this same way that the human capability country and the USA are now offering" can be less effective when the brain the public microcomputer hobby kits is not coupled to five good senses and for as little as £40 and microprocessors are being incorporated in domestic healthy muscular action. equipment and motor cars. Chess At the present stage of developsets with a microcomputers able to ment the applications likely to progress play the game against human opponents fastest are those which are simple enough are available and a 'Speak and Spell' to have proprietary application programs toy has been developed using a microbuilt into the chips via the final maskprocessor chip to synthesize human ing in their manufacture. On economic speech. Although some of these grounds, this implies at least 500 examples are not directly relevant to • identical chips being required. instrumentation it can be inferred This figure of 500 might be regarded that a huge domestic demand will as low, but it can be justified on the bring cost reduction benefits to microgrounds that instrument manufacture computer hardware which will have a will be able to achieve major cost tremendous influence on instrument reductions by the employment of technology. Even today a microprocessol chips, so that it will be economically feasible for them to pay prices which chip can be bought for only £3. The UK are much higher than those which Government, having realized the apply to chips for the mass markets. potential impact of microprocessors The Instrument Industry will on industry, have offered large sums Table 1. Applications of microelectronics in flow measurement
Multiplex signals from transducers to simplify cabling Control interrogation of transducers and data buses Data Presentation : Data processing Control of VDUs and other displays Control of priorities Alarms: Provide alarm logic
Transmission:
Control :
Control alarm devices Incorporation in discrete digital controllers Incorporation in distributed control systems Provide automatic tuning of controllers Flexible programmable controllers
Diagnostics Actuators: Transducers:
Signal conditioners Control speed, position and safety routines Variable characterization and loop gain adjustment Self-calibration and fault diagnosis
Adjustment for drift Linearization
Improvement in accuracy and repeatability by eliminating systematic errors Data processing of data from 2 or more transducers to compute, for example, mass.flow or flow measurement by cross correlation Analogue to tligital transmission of data and vice versa Control of analysers particularly mass spectrometers and gas chromatographs
microprocessors and microsystems
require an inqreasing number of microcircuits. These could take the form of custom-designed LSl or microprocessors. A problem likely to face the instrument manufacturer is that of obtaining the cooperation of the semicopductor manufacturers who seem to be predisposed to the large scale consumer markets. Using CAD techniques, custom-designed microcircuits selling at say £25 in lots of 500 units could be developed and produced which would be an economic proposition for both the semiconductor manufacturer and the instrument manufacturer who could use them to reduce his costs by a substantial margin. However the apparent lack of interest by the semiconductor manufacturer, together with the dependence on one supplier, the rapid changes in techno!ogy and the inflexibility of this approach, will influence instrument designers in the future to meet their applicational requirement through software and microprocessors. Another possible line of development where the amount of data processing is small and the number of input/output channels is relatively few, is that of the single chip microcomputer. This would cover in a single hardware design, a wider range of applications than that of the custom built LSI, through the use of software written into read only memory. The software problems, that is the cost, the manpower shortage and the lack of standardization, will dominate the microprocessor scene for many years and will control growth. Some relief will be obtained by the availability of more powerful computing facilities, bigger and cheaper memories and the adoption of hardware in place of software. Magnetic bubble memories Which are nonvolatile are becoming available. It has been forecast that by 1982 a bubble memory capable of storing .1 Mbit and having a i/~s response time will be avaiiable on a single chip 25 mm square and will cost about £50. Although the greatest excitement has been generated by the advent of the microprocessor, nevertheless the technical achievement of cheap large scale memories deserves equal attention.
vol 3 no I 0 december 79
It is fairly certain that microcomputers will be oversold and misapplied and give rise to some disitlusion~nent in common with every new technical tide. This disenchantment will not be due to lack of reliability or performance, but to misapplication, overselling, inadequate memory, inadequate word length, rapid ob~;olescence and shoddy software, all born from hasty optimism and the desire to make a quick return. It has to be remembered that a microcomputer is relatively more difficult to program than a minicomputer; also a microcomputer is only as capable as an ordinary mainframe computer in performing tasks. The revolution brought about by the microprocessor is that it can perform these tasks more cheaply and flexibly. In a recent Frost and Sullivan report it was estimated that by 1981 the European process control market will require £20 million worth of microcomputers and this figure will rise to £100 million by 1986.
development and sale can be confidently predicted. Three-term analogue electronic controllers will be required in economic production quantities for about seven years, but it would be hazardous to forecast a longer life now that a microprocessor single loop controller could possibly be produced at a cost comparable with the analogue version. In any case the single loop controller will be restricted to small plants and for back-up purposes until the cost of alternatives can be reduced sufficiently for manufacturers to employ a high degree of redundancy. Indeed, a major advance expected in the future for all microcomputer-I~ased equipment is the improvement in reliability through redundancy and se[f diagnostics. It seems probable that manufacturers will take advantage of the continuously reducing prices of electronic components to improve performance and reliability for the same product price rather than maintain previous standards at a lower price.
CONTROLLERS
Analogue and digital controller~ These controllers with their valuable single loop integrity will continue to be used for many years to come but will give way to integrated systems combining distributed control functions with computer supervision, and advanced forms of communication systems. Microprocessor-based digital controllers will compete strongly with analogue types as costs become comparable and as the users recognize the advantages of the digital controller in terms of accuracy, flexibility, adaptibility and ease of interfacing with other digital equipment.
Programmable controllers One of the largest growth areas in control is the programmable controller. World sales of these controllers are estimated to be currently £100 million. They are applied 60% to manufacturing and 40% to process industries and for the last two yea'rs have shown a growth rate of 100% per annum. The basics of a programmable controller can now be purchased very cheaply as a single.silicon chip, so a further escalation of
MEASUREMENT
.
Future developments in this area will be largely influenced by the application of more silicon technology, particularly the application of microprocessors. The benefits will apply more to the secondary aspects of measurement than to the primary.
Flow Flow has enjoyed two decades of major development and there is now a wide choice of flowmeters for most applications. It is unlikely that a radically new technique of industrial flow measurement will appear in the next ten years. But improvement of existing types is badly needed including the application of electronics to improve accuracy standards, Table 2 gives some of the likely areas of improvement.
Pressure Developments in pressure measurement seem destined for the next decade to be related to diffused or ion implanted silicon diaphragms with or without integral amplifier~. Longer term, optical
475
measurement of diaphragm movement by optical fibres could come about (see Transducers).
Tern peratu re No revolutionary changes can be expected in temperature sensing materials but some minor developments will occur involving the use of semiconductor materials as temperature sensors. Fibre optics pyrometers have recently been developed and will play a useful role in certain difficult applications of temperature measurement.
Level Ultrasonic techniques have a good potential for this type of measurement for liquids and solids and will probably show a higher growth rate than other techniques.
Analytical A great deal could happen in this field but it is unlikely that there will be any single large scale development. Infrared, gas chromatography and mass spectrometry will benefit from development expenditure. New developments can be expected in specific ion measurement in liquids involving selective membranes; for gases, Semiconductor sensors could make a useful contribution but their performance to date has not been very encouraging.
Table 2. Areas of likely improvement in flowmeters Pressure difference Research being undertaken in various parts of the world should refine devices available knowledge on coefficient data and on the effects of asymmetrical flow profiles. Pressure differences will be measured more accurately by transducers (q.v.). Electromagnetic Direct current pulse excitation in various forms will continue to be developed and refined, but alternating current excitation could be improved to compete with the newer systems being offered. Improved electronic circuitry will permit the application of these meters to nonconducting fluids. Vortex This still promises to be the meter of the future but both electronics and vortex detection must be improved. Other types of flowmeters based on other forms of hydrodynamic instability will appear on the market but, because they rely on similar principles to that of the vortex meter, will experience similar limitations in accuracy and performance. Cross correlation One commercial meter has appeared on the US market, Severalmore will appear in the next five years for application to some of the more difficult two-phase fluids. The correlator will almost certainly be microprocessorbased. Turbine This type of flowmeter still has a long life ahead of it but development work will be aimed at refinement rather than at fundamental change. Improved pick-offs could reduce cost and improve performance. Mass flow Microprocessors will increasingly be used in the various systems of mass flow measurement both for calculating corrections for a range of variables and for controlling the input of data. Ultrasonic This type of meter has not enjoyed a high reputation in the past although it has been offered in the market for 20 years or more. It is possible that it could emerge as a main competitor to most other types.
problem. The basic idea is simple; to use light conducted down a fibre to measure the strain or displacement o f a transducer diaphragm. The light would then be modulated and returned to the control room for interpretation. For the moment the ideal light modulator for this application does not exist but when a solution is found it will offer an exciting o p p o r t u n i t y to the manufactu rer.
TRANSDUCERS In the last five years the market has swung away from force balance designs and has welcomed microdisplacement types of which the two most important contenders are those o f the silicon strain gauge and the variable capacity. A t the present 'state of the art' the variable capacity system is well established and is likely to persist in the market for at least a decade. However, the silicon strain gauge approach is attractive and is linked to a rapidly advancing field of technology, at least as far as the silicon chip is concerned. In the interest of plant safety, there will be a demand for 'passive~ transducers, that is, transducers consuming no power in situ. Interest is turning to fibre optics to solve this
476
TELEMETRY Economic, efficient and reliable telemetry will be one of the main contributors to modern developments in the integrated control of process plants. In the past, measurement and control instrumentation was necessarily distributed around the plant. As improved short distance transmission signals evolved, control became centralized in one control room. Recently the growth in size and c o m p l e x i t y of plants and the necessity for a more hierarchical organization of information and control is again leading to a distributed control system, but this is not a complete reversal. Modern systems have a facility for transmission of data which could scarcely be envisaged 30 years ago enabling a
wide distribution of equipment to be made w i t h o u t loss of essential information at any point. Transfer of large amounts of data can only be carried out economically in a digital mode. The transfer medium can be multicore cable, coaxial cable, microwave or optical fibre. Technological developments o f the last two will have an increasing impact on process control.
DISPLAYS The increasing amounts o f data required by some modern processes have made the instrument panels of ten years ago quite impracticable. Various alternatives have been introduced to overcome the difficulty, the most successful being the VDU in the form of a black and white or colour television type screen. In general terms, one V D U can replace 15 m of conventional panel. The V D U and its associated controls become the interface between the operator and the process. There will be constant pressure from the user to improve this interface still further, particularly as far as the display feature is concerned. New devices employing plasma discharge panels are coming on to the market which, though simpler, less bulky and probably cheaper than the
microprocessors and microsystems
television type of VDU, are no more pleasant to look at. A more attractive possibility could be projection type liquid crystal displays but these are a long way off. 'Nevertheless for other instrument displays, liquid crystals could be the main challenger and the instrument industry could reap the benefit of an increasing domestic demand for cheap displays for watches and calculators.
subject is outside the scope of this article as the current use for robots is for physical jobs involving health hazards and repetitive tasks - welding being a typical example. Nevertheless robots are mentioned because ultimately their development will interact with the requirements of process control.
U LT RASON ICS
In general terms the development of process control instrumentation has worked towards the long term goal of equalling or supassing the capabilities of the human being in terms of the senses (hearing, sight, touch, taste, smell) calculation, memory and logic (the brain) and in physical actuation (the muscles). In spite of the tremendous advances of science the gaps between human capability and instrument capability are enormous in some areas, and it is these gaps which provide the pointers to.the direction of future developments. For example, pattern recognition by computer is very weak by human standards; chemical analysis is cumbersome and slow compared with the speed arid sensitivity of taste and smell. Photoelectric and optical systems do not have the dynamic range of the eye, and microphones do not have the dynamic range and noise discrimination of the ear (1030 to 1). As far as the brain is concerned
Now that ultrasonic transducers and associated electronics are well developed and available at economic prices, the application of ultrasonic techniques is likely to accelerate in the instrument industry. Apart from the known uses of these techniques in flow, chemical concentration, level and electrode cleaning, there could be many other applications such as the accurate rneasurement of displacement, consistency, gas analysis and tern peratu re.
ROBOT~ To many people, the term 'robots' conjures up thoughts of science fiction, but there are over 3000 robots operating in the USA and probably as many in Japan. The UK is considered to be ten years behind Japan in 'robitics' and the Government has now started to become concerned and is sending a study team to Japan. The
Z8 prototype device Zilog has announced the availability of the Z8-02 MPD, a 64-pin version of the Z8 single chip microcomputer. It is specifically intended for prototype development where program changes are expected as development progresses. The 64-pin device, a ROMless version of the standard 40-pin ZS, provides 12 ROM address lines, 8 ROM data lines and necessary control signals. External program memory in the form of EPROMs, bipolar PROMs or RAMs can be used to interface to the Z8-02 MPD in developing the first 2 kbyte of program. When the final program has been established, large volume applications would normally
vol 3 no 10 december 79
CONCLUSION
use mask programmed ROM versions of the standard 40-pin Z8 (Z8-01). Factory samples of the Z8-02 MPD prototype device are now available in ceramic quad-in-line package at around £$5, and mask programmed Z8-01s should be available in quantity from early next year.
Zilog's 64-pin Z8 for prototype development
computers have a long way to go; the brain normally has ten thousand million neurons for storage of information with an access time of ls (sometimes a little longer.). Each neuron can communicate internally with 5000 other neurons providing facilities for an almost infinite variety of memory patterns. The bubble memory of 1982 will be able to store 1 Mbit of information (only accessible externally) in a 25 mm square chip. For comparison, 2000 of these bubble memories would be needed to store the contents of the Encyclopedia Britannica. The brain also has built into it a vast amount of redundancy in order to ensure that false signals and noise do not destroy the value of the information received. Thus, there is a long way to go to bring computers up to the memory capacity of the human brain. Sometimes in the next 50 years there could be a breakthrough in reducing some of the large gaps between human capabilities and those of instruments but the probability is that it will involve the development engineer in a further science, such as biology or physiochemistry, for there is no doubt that biological sensors have a specificity and sensitivity greatly exceeding those made of metal, plastic and silicon. This information first appeared in Kent Technical Review No 24.
Changing names Computer Marketing , and Micro Software Systems which it took over less than three months ago, will in future be known as Comma Computers Limited. Comma sells the full range of Newbury terminals and systems, Tally printers, Anderson-J acobson acoustic couplers, Diablo printers, Sigma graphics and the Princeton 8500M high resolution graphics terminals. Comma also manufacturers and markets a range of Comma Systems microcomputers based on North Star Horizon and on Digital Equipment VO3 and TO3 processors - the Comma Leo, Aquarius, Capricorn, Libra and Gemini.
477
of taxpayers' money to encourage their manufacture and application.
R e g M e d l o c k o f K e n t Instruments examines p o s s i b l e areas o f advancement f o r i n s t r u m e n t technology, particularly in improving t h e e f f i c i e n c y a n d s a f e t y o f t h e process industries
Microelectronic
The process industries have a tradition of conservatism towards the adoption of new products whose performance, reliability and back-up resources are still to be proved. Furthermore, through replacement, retrofit and user familiarity, the life of older products is extended, sometimes to Obsolescence, thereby retarding the acceptance of newer developments. The effect of this is that the prophets of technology tend to overestimate progress in the first five years but underestimate in the following 20 years. M ICROE L ECTRON
ICS
There is however, a unanimity of opinion amongst forecasters that the instrument industry will be dominated by microelectronics for the foreseeable future. In this article microelectronics covers microprocessors, memories, silicon chip devices, integrated circuits, LSI and VLSI. The news media frequently refer to microprocessors when they really mean microcomputers or even the broader field of microelectronics. The precise boundary between microprocessors and microcomputers is difficult to establish as it dep'ends on what has been included on the microprocessor chip. Normally a microprocessor is a component and a microcomputer is an assembled kit on a printed circuit board comprising one or more microprocessor units and additional components such as memories. The growth of computers in industrial control was slow I~o start but has begun to reach explosive proportions. It has been estimated that, since the first computer control applications in 1960, half a million central processors have been installed (not just on process control of course). The growth rato has averaged 67% per annum and it is believed that 240 000 processors
474
applications
Microcircuitry including microprocessors and other microcircuit chips will be applied to instruments in many ways. Table 1 shows a n u ~ of.e~amples. In the excitement of microprocessor were added in ] 978 and 340 000 will application it is sometimes forgotten be added in 1979. that usage will be limited by problems Microcomputers look as if they associated with interfacing, software will invade all industry as well as the production and mechanisms in the domestic market. Retail shops in this same way that the human capability country and the USA are now offering" can be less effective when the brain the public microcomputer hobby kits is not coupled to five good senses and for as little as £40 and microprocessors are being incorporated in domestic healthy muscular action. equipment and motor cars. Chess At the present stage of developsets with a microcomputers able to ment the applications likely to progress play the game against human opponents fastest are those which are simple enough are available and a 'Speak and Spell' to have proprietary application programs toy has been developed using a microbuilt into the chips via the final maskprocessor chip to synthesize human ing in their manufacture. On economic speech. Although some of these grounds, this implies at least 500 examples are not directly relevant to • identical chips being required. instrumentation it can be inferred This figure of 500 might be regarded that a huge domestic demand will as low, but it can be justified on the bring cost reduction benefits to microgrounds that instrument manufacture computer hardware which will have a will be able to achieve major cost tremendous influence on instrument reductions by the employment of technology. Even today a microprocessol chips, so that it will be economically feasible for them to pay prices which chip can be bought for only £3. The UK are much higher than those which Government, having realized the apply to chips for the mass markets. potential impact of microprocessors The Instrument Industry will on industry, have offered large sums Table 1. Applications of microelectronics in flow measurement
Multiplex signals from transducers to simplify cabling Control interrogation of transducers and data buses Data Presentation : Data processing Control of VDUs and other displays Control of priorities Alarms: Provide alarm logic
Transmission:
Control :
Control alarm devices Incorporation in discrete digital controllers Incorporation in distributed control systems Provide automatic tuning of controllers Flexible programmable controllers
Diagnostics Actuators: Transducers:
Signal conditioners Control speed, position and safety routines Variable characterization and loop gain adjustment Self-calibration and fault diagnosis
Adjustment for drift Linearization
Improvement in accuracy and repeatability by eliminating systematic errors Data processing of data from 2 or more transducers to compute, for example, mass.flow or flow measurement by cross correlation Analogue to tligital transmission of data and vice versa Control of analysers particularly mass spectrometers and gas chromatographs
microprocessors and microsystems
require an inqreasing number of microcircuits. These could take the form of custom-designed LSl or microprocessors. A problem likely to face the instrument manufacturer is that of obtaining the cooperation of the semicopductor manufacturers who seem to be predisposed to the large scale consumer markets. Using CAD techniques, custom-designed microcircuits selling at say £25 in lots of 500 units could be developed and produced which would be an economic proposition for both the semiconductor manufacturer and the instrument manufacturer who could use them to reduce his costs by a substantial margin. However the apparent lack of interest by the semiconductor manufacturer, together with the dependence on one supplier, the rapid changes in techno!ogy and the inflexibility of this approach, will influence instrument designers in the future to meet their applicational requirement through software and microprocessors. Another possible line of development where the amount of data processing is small and the number of input/output channels is relatively few, is that of the single chip microcomputer. This would cover in a single hardware design, a wider range of applications than that of the custom built LSI, through the use of software written into read only memory. The software problems, that is the cost, the manpower shortage and the lack of standardization, will dominate the microprocessor scene for many years and will control growth. Some relief will be obtained by the availability of more powerful computing facilities, bigger and cheaper memories and the adoption of hardware in place of software. Magnetic bubble memories Which are nonvolatile are becoming available. It has been forecast that by 1982 a bubble memory capable of storing .1 Mbit and having a i/~s response time will be avaiiable on a single chip 25 mm square and will cost about £50. Although the greatest excitement has been generated by the advent of the microprocessor, nevertheless the technical achievement of cheap large scale memories deserves equal attention.
vol 3 no I 0 december 79
It is fairly certain that microcomputers will be oversold and misapplied and give rise to some disitlusion~nent in common with every new technical tide. This disenchantment will not be due to lack of reliability or performance, but to misapplication, overselling, inadequate memory, inadequate word length, rapid ob~;olescence and shoddy software, all born from hasty optimism and the desire to make a quick return. It has to be remembered that a microcomputer is relatively more difficult to program than a minicomputer; also a microcomputer is only as capable as an ordinary mainframe computer in performing tasks. The revolution brought about by the microprocessor is that it can perform these tasks more cheaply and flexibly. In a recent Frost and Sullivan report it was estimated that by 1981 the European process control market will require £20 million worth of microcomputers and this figure will rise to £100 million by 1986.
development and sale can be confidently predicted. Three-term analogue electronic controllers will be required in economic production quantities for about seven years, but it would be hazardous to forecast a longer life now that a microprocessor single loop controller could possibly be produced at a cost comparable with the analogue version. In any case the single loop controller will be restricted to small plants and for back-up purposes until the cost of alternatives can be reduced sufficiently for manufacturers to employ a high degree of redundancy. Indeed, a major advance expected in the future for all microcomputer-I~ased equipment is the improvement in reliability through redundancy and se[f diagnostics. It seems probable that manufacturers will take advantage of the continuously reducing prices of electronic components to improve performance and reliability for the same product price rather than maintain previous standards at a lower price.
CONTROLLERS
Analogue and digital controller~ These controllers with their valuable single loop integrity will continue to be used for many years to come but will give way to integrated systems combining distributed control functions with computer supervision, and advanced forms of communication systems. Microprocessor-based digital controllers will compete strongly with analogue types as costs become comparable and as the users recognize the advantages of the digital controller in terms of accuracy, flexibility, adaptibility and ease of interfacing with other digital equipment.
Programmable controllers One of the largest growth areas in control is the programmable controller. World sales of these controllers are estimated to be currently £100 million. They are applied 60% to manufacturing and 40% to process industries and for the last two yea'rs have shown a growth rate of 100% per annum. The basics of a programmable controller can now be purchased very cheaply as a single.silicon chip, so a further escalation of
MEASUREMENT
.
Future developments in this area will be largely influenced by the application of more silicon technology, particularly the application of microprocessors. The benefits will apply more to the secondary aspects of measurement than to the primary.
Flow Flow has enjoyed two decades of major development and there is now a wide choice of flowmeters for most applications. It is unlikely that a radically new technique of industrial flow measurement will appear in the next ten years. But improvement of existing types is badly needed including the application of electronics to improve accuracy standards, Table 2 gives some of the likely areas of improvement.
Pressure Developments in pressure measurement seem destined for the next decade to be related to diffused or ion implanted silicon diaphragms with or without integral amplifier~. Longer term, optical
475
measurement of diaphragm movement by optical fibres could come about (see Transducers).
Tern peratu re No revolutionary changes can be expected in temperature sensing materials but some minor developments will occur involving the use of semiconductor materials as temperature sensors. Fibre optics pyrometers have recently been developed and will play a useful role in certain difficult applications of temperature measurement.
Level Ultrasonic techniques have a good potential for this type of measurement for liquids and solids and will probably show a higher growth rate than other techniques.
Analytical A great deal could happen in this field but it is unlikely that there will be any single large scale development. Infrared, gas chromatography and mass spectrometry will benefit from development expenditure. New developments can be expected in specific ion measurement in liquids involving selective membranes; for gases, Semiconductor sensors could make a useful contribution but their performance to date has not been very encouraging.
Table 2. Areas of likely improvement in flowmeters Pressure difference Research being undertaken in various parts of the world should refine devices available knowledge on coefficient data and on the effects of asymmetrical flow profiles. Pressure differences will be measured more accurately by transducers (q.v.). Electromagnetic Direct current pulse excitation in various forms will continue to be developed and refined, but alternating current excitation could be improved to compete with the newer systems being offered. Improved electronic circuitry will permit the application of these meters to nonconducting fluids. Vortex This still promises to be the meter of the future but both electronics and vortex detection must be improved. Other types of flowmeters based on other forms of hydrodynamic instability will appear on the market but, because they rely on similar principles to that of the vortex meter, will experience similar limitations in accuracy and performance. Cross correlation One commercial meter has appeared on the US market, Severalmore will appear in the next five years for application to some of the more difficult two-phase fluids. The correlator will almost certainly be microprocessorbased. Turbine This type of flowmeter still has a long life ahead of it but development work will be aimed at refinement rather than at fundamental change. Improved pick-offs could reduce cost and improve performance. Mass flow Microprocessors will increasingly be used in the various systems of mass flow measurement both for calculating corrections for a range of variables and for controlling the input of data. Ultrasonic This type of meter has not enjoyed a high reputation in the past although it has been offered in the market for 20 years or more. It is possible that it could emerge as a main competitor to most other types.
problem. The basic idea is simple; to use light conducted down a fibre to measure the strain or displacement o f a transducer diaphragm. The light would then be modulated and returned to the control room for interpretation. For the moment the ideal light modulator for this application does not exist but when a solution is found it will offer an exciting o p p o r t u n i t y to the manufactu rer.
TRANSDUCERS In the last five years the market has swung away from force balance designs and has welcomed microdisplacement types of which the two most important contenders are those o f the silicon strain gauge and the variable capacity. A t the present 'state of the art' the variable capacity system is well established and is likely to persist in the market for at least a decade. However, the silicon strain gauge approach is attractive and is linked to a rapidly advancing field of technology, at least as far as the silicon chip is concerned. In the interest of plant safety, there will be a demand for 'passive~ transducers, that is, transducers consuming no power in situ. Interest is turning to fibre optics to solve this
476
TELEMETRY Economic, efficient and reliable telemetry will be one of the main contributors to modern developments in the integrated control of process plants. In the past, measurement and control instrumentation was necessarily distributed around the plant. As improved short distance transmission signals evolved, control became centralized in one control room. Recently the growth in size and c o m p l e x i t y of plants and the necessity for a more hierarchical organization of information and control is again leading to a distributed control system, but this is not a complete reversal. Modern systems have a facility for transmission of data which could scarcely be envisaged 30 years ago enabling a
wide distribution of equipment to be made w i t h o u t loss of essential information at any point. Transfer of large amounts of data can only be carried out economically in a digital mode. The transfer medium can be multicore cable, coaxial cable, microwave or optical fibre. Technological developments o f the last two will have an increasing impact on process control.
DISPLAYS The increasing amounts o f data required by some modern processes have made the instrument panels of ten years ago quite impracticable. Various alternatives have been introduced to overcome the difficulty, the most successful being the VDU in the form of a black and white or colour television type screen. In general terms, one V D U can replace 15 m of conventional panel. The V D U and its associated controls become the interface between the operator and the process. There will be constant pressure from the user to improve this interface still further, particularly as far as the display feature is concerned. New devices employing plasma discharge panels are coming on to the market which, though simpler, less bulky and probably cheaper than the
microprocessors and microsystems
television type of VDU, are no more pleasant to look at. A more attractive possibility could be projection type liquid crystal displays but these are a long way off. 'Nevertheless for other instrument displays, liquid crystals could be the main challenger and the instrument industry could reap the benefit of an increasing domestic demand for cheap displays for watches and calculators.
subject is outside the scope of this article as the current use for robots is for physical jobs involving health hazards and repetitive tasks - welding being a typical example. Nevertheless robots are mentioned because ultimately their development will interact with the requirements of process control.
U LT RASON ICS
In general terms the development of process control instrumentation has worked towards the long term goal of equalling or supassing the capabilities of the human being in terms of the senses (hearing, sight, touch, taste, smell) calculation, memory and logic (the brain) and in physical actuation (the muscles). In spite of the tremendous advances of science the gaps between human capability and instrument capability are enormous in some areas, and it is these gaps which provide the pointers to.the direction of future developments. For example, pattern recognition by computer is very weak by human standards; chemical analysis is cumbersome and slow compared with the speed arid sensitivity of taste and smell. Photoelectric and optical systems do not have the dynamic range of the eye, and microphones do not have the dynamic range and noise discrimination of the ear (1030 to 1). As far as the brain is concerned
Now that ultrasonic transducers and associated electronics are well developed and available at economic prices, the application of ultrasonic techniques is likely to accelerate in the instrument industry. Apart from the known uses of these techniques in flow, chemical concentration, level and electrode cleaning, there could be many other applications such as the accurate rneasurement of displacement, consistency, gas analysis and tern peratu re.
ROBOT~ To many people, the term 'robots' conjures up thoughts of science fiction, but there are over 3000 robots operating in the USA and probably as many in Japan. The UK is considered to be ten years behind Japan in 'robitics' and the Government has now started to become concerned and is sending a study team to Japan. The
Z8 prototype device Zilog has announced the availability of the Z8-02 MPD, a 64-pin version of the Z8 single chip microcomputer. It is specifically intended for prototype development where program changes are expected as development progresses. The 64-pin device, a ROMless version of the standard 40-pin ZS, provides 12 ROM address lines, 8 ROM data lines and necessary control signals. External program memory in the form of EPROMs, bipolar PROMs or RAMs can be used to interface to the Z8-02 MPD in developing the first 2 kbyte of program. When the final program has been established, large volume applications would normally
vol 3 no 10 december 79
CONCLUSION
use mask programmed ROM versions of the standard 40-pin Z8 (Z8-01). Factory samples of the Z8-02 MPD prototype device are now available in ceramic quad-in-line package at around £$5, and mask programmed Z8-01s should be available in quantity from early next year.
Zilog's 64-pin Z8 for prototype development
computers have a long way to go; the brain normally has ten thousand million neurons for storage of information with an access time of ls (sometimes a little longer.). Each neuron can communicate internally with 5000 other neurons providing facilities for an almost infinite variety of memory patterns. The bubble memory of 1982 will be able to store 1 Mbit of information (only accessible externally) in a 25 mm square chip. For comparison, 2000 of these bubble memories would be needed to store the contents of the Encyclopedia Britannica. The brain also has built into it a vast amount of redundancy in order to ensure that false signals and noise do not destroy the value of the information received. Thus, there is a long way to go to bring computers up to the memory capacity of the human brain. Sometimes in the next 50 years there could be a breakthrough in reducing some of the large gaps between human capabilities and those of instruments but the probability is that it will involve the development engineer in a further science, such as biology or physiochemistry, for there is no doubt that biological sensors have a specificity and sensitivity greatly exceeding those made of metal, plastic and silicon. This information first appeared in Kent Technical Review No 24.
Changing names Computer Marketing , and Micro Software Systems which it took over less than three months ago, will in future be known as Comma Computers Limited. Comma sells the full range of Newbury terminals and systems, Tally printers, Anderson-J acobson acoustic couplers, Diablo printers, Sigma graphics and the Princeton 8500M high resolution graphics terminals. Comma also manufacturers and markets a range of Comma Systems microcomputers based on North Star Horizon and on Digital Equipment VO3 and TO3 processors - the Comma Leo, Aquarius, Capricorn, Libra and Gemini.
477