The programmable automaton, a universe of possibilities in constant evolution

Was Dick Morly, at present known like the father of the PLCs, the one who was involved in the development of this first PLC for GE, known like Modicon (Modulate Digital CONtroller). The aim and design of a PLC does not be distant so much of the of a conventional PC. Both are formed mainly by entrances, exits, memory and a CPU. And of the same way, his elementary aim is, in front of the conmutación in one or miscellaneous of his entrances, realise logical calculations and, like result, activate or desactivar one or several exits. However, there is a fundamental appearance where a PLC differs of the computers ofimáticos: the robustness. East was, precisely, one of the fundamental appearances that GE looked for for his developments: a very reliable device, able to operate in the complicated industrial conditions, with ranks of temperature, humidity and feeding extended, with a high resistance to the vibrations and to the impacts and, of course, with a high immunity to the electromagnetic noise. Another substantial difference is the number of And/S (entrances and/or exits) that it is able to handle a PLC. Unlike a conventional PC, any PLC current part of cientos of And/S until arriving to the thousands that can manage a PLC current of high range.

During these more than 40 years, the PLCs have evolved in a lot of appearances and, if in his starts his fundamental mission was to substitute to the logic wired up, nowadays, with the technological advances and the big reduction of costs, his aim arrives much more there that the one to do logical calculations.

Like this, in addition to the control of the discreet logic, the control of continuous processes with analog signals or of temperature was one of the first incorporations to the world of the PLC.

But has been from the year 2000 when the world of the PLC has suffered a greater advance. With the reduction of costs in the electronic equipment more complex and in the electronics equipment dealers equipment dealers of power, arose the specific controllers for artificial vision, for ‘Motion Control', hygiene, etc.

With all these controllers, apart from the PLC, arose the need imperiosa of the integration: with different networks of communication, different software of configuration, different programming languages, etc.

Nowadays, in the eagerness to simplify the integration of all these systems, arises the concept of MAC (Machine Automation Controller) or PAC (Programmable Automation Controller), where the PLC evolves to bear many of the before commented disciplines. So that from an only software, with a leguaje and interface common and without any problem of interconnection, an only controller can do charge of the automation of a machine or a line of whole production.

One of the problems that appeared with the evolution and commercialisation of the different models of PLC, is the fragmentation in the programming languages used in each one of the equipment, stressed besides by own languages of the different marks.

The language with which began all was the ‘ladder'. This language, invented to represent of a clear way the logic realised with relays, and whose diagram, with the different connections and contacts, remembers to the of some ladders (‘ladder' in English) with his different ladders (‘rungs').

However, with the inclusion of new functionality in the PLCs were appearing new languages, more adapted to realise other tasks, like arithmetical calculations complexes, conditional sentences or that allowed a greater reuse of the code.

Like this, to finals of 1993, the IEC published the first review of the standard 61131, that in his third section defines the standard programming languages of a PLC. At present they are the following:

• Ladder Diagram (LD, diafragma of ladders)
• Function Block Diagram (FBD, diagram of blocks of function)
• Structured Text (ST, text structured)
• Instruction List (IL, list of instructions)
• Sequential Function Chart (SFC, chart of sequential functions)

This provides a very important fan of possibilities in the programming of the current controllers. Even, some manufacturers allow to mix miscellaneous of them in a same program, doing that the task of development was a lot of simpler.

On the other hand, the norm IEC previous, no only limits to define the programming languages, but also the types of data and some own appearances of the operation of the own controller: gestión of data using variable, use of tasks, etc.

This standardisation, beside others, does that of a very simple way, can use diverse controllers without having to learn any new language or form to program.

If we throw a glimpse to the current controllers (the MAC or PAC) that commented before, can observe like the architecture has happened to use CPUs own, of type ASIC to standard microprocessors of market; usually AMD or Intel.

This brings achieve important advances in the possibilities and speed of the controllers, what to his time has made possible the integration of more disciplines in an only team.

Like this, in any of the most powerful equipment of the market, can find, for example, microprocessors Intel Atom, to 1,6 GHz, that provide a speed of process never before sight in industrial controllers.

This speed also depends to a large extent of the software that have to bear this CPU. Like this, it is usual to find controllers that execute Windows CE or Windows Embedded that, of usual form, need more resources to work that other Operating systems, called ‘Hard Real Time YOU', oriented to the industrial control and that guarantee a repetitividad and reliability so high that they are the elected for the control in other sectors, like the aerospace, the one of defence, in medicine, etc.

Examples of this type of ‘Hard Real Time YOU' are QNX or the Microware YOU-9.

On the operating system, the controllers MAC or PAC execute the software ‘runtime' of the own manufacturer. This software ‘runtime' is used to to be modulate and is the attendant to execute the logical sequences (module PLC), of the control of movement (module of ‘Motion Control'), of the control of robots (‘kinematics'), CNC, etc.

Usually exists a main module that is used to to be the one of PLC, that commissions of the logic and realises the timely calls to the rest of modules when in program of user requires of these modules.

Imagine a simple program of a machine of cutting to the flight, attendant to cut each one of the containers that will happen to the following stage inside the line of production and where the exact point of cut will come given by a mark in the own container.

In the strip transportadora, by where displaces the film (throw continuous of containers, still without cutting), will place a fotocélula, to a determinate distance of the blade. This fotocélula will be the attendant to detect the present marks in the film so that the blade arrive to move to the suitable speed and arrive to the point of cut to the same speed exactly of the strip transportadora.

In this example, if the module of PLC is the main, will commission to initiate the execution of program and in each one of the movements, will do an application to the module of ‘Motion Control'. These movements are very very clear-cut, as at present almost all the manufacturers use the standard PLCopen (to his time, based in the IEC 61131-3) for the control of movement and each one of the instructions that do use of this module begin by MC. For example:

• Initialise the axles and that the ‘drives' deliver power to the motors
• Realise the research of origin in the axis of the blade
• Begin the movement of the strip, linear and endless in the axis of the strip
• And once detected the mark, establish the necessary synchronisation between the two axles so that the rotatory blade reach the mark in the just moment
• Turn to the origin of the axis of the blade

Incidentally, the main module, in this case the one of PLC, also is used to to be attendant of other tasks, as it is the soda of the And/S and the attention to peripherals (attention to ports USB, Ethernet, access to cards by heart, etc.).

Commented some lines before that IEC 61131-3 defined no only the programming languages, but it arrived to specify the architecture of operation of the controllers.

In this way, the controllers based in this standard have all something in common: a model of execution by tasks.

Exist two models of execution of tasks:

• Non-Yielding (improductivo)
• Yielding (productive)

In the first of them, a task has to expect to that another finish to be able to execute. In the second, establishes a system of priorities, so that a task with a greater priority can interrupt to another and begin his execution. The task interrupted will restart his execution when it finish the execution of the tasks of greater priority.

This last system offers a greater performance, to the pair that allows to guarantee the determinism of the system; that is to say, that a same sequence of a determinate priority execute always in the same time, independently of the load (number of sequences in execution in this moment) of the system. However, this last system of gestión of tasks also is more complex and requires of an Operating system of real time, by what can not be present in all the current controllers.

In this productive system exist tasks of diverse priority, as we said, although they also can exist tasks of evento, of attention to peripherals or, simply, of internal processes.

Interest us those tasks to which can them assign programs of user. This model, defined by the norm IEC, allows to assign multiple programs to each one of the tasks of different priority.

This comports that the programs assigned to a task of greater priority always will execute before that the assigned to a task of inferior priority. And inside this model, the programs assigned to a same task will execute cíclicamente, one to continuation of another. This order usually can establish in the software of programming.

So important like a good allocation of tasks, is the allocation of the soda of And/S, since a good adjust of the priorities and the allocations will allow to carry the cycles of the machine to the minimum.

This soda of the And/S, when it treats of units of And/S local does not represent any problem. But however, when it treats of units of And/S distributed or have to effect a control of movement of the maximum precision, is necessary to support in any of the best networks of communications of the market, in which connect these servo accionamientos, And/S and another periphery.

In the actuality exist multiple networks that allow to integrate And/S, servos and other a lot of devices. However, if we look carefully the technical specifications, real results and the expansion that is suffering in the market, can speak of EtherCAT like the network that stands out on the other, to level of machine.

A characteristic of EtherCAT is that it uses wire standard Ethernet to realise the wired up between the controller and the different nodes of the network. At the same time, each node has of two ports: one of entrance and another of exit.

Of this way, can build a network ‘daisy-chain' without any additional hardware and a totally standard wire and economic. On the other hand, EtherCAT is flexible and, if in a moment needs mount another topology of network, distinct of the one of chain, exist bifurcadores in the market that allow us create any structure in tree.

But the main characteristic of this network does not find in the ease of wired up, but in the rapidity. The operation, in spite of using wire standard Ethernet, are far from to be the one of a conventional Ethernet. And it is that in EtherCAT, the teacher launches a plot of fixed size (determined by the number of nodes and the information to exchange by each one of them) and arrives to the first of the nodes. This, reads the information that needs and writes another information in the plot, in his space reserved for this.

To continuation, this node sends this plot modified to the following, and east goes back to read and write data in the space that has reserved for this. When arriving to the last node, this gives back the plot, already with all the data, to the teacher; always working in a network ‘full-duplex', by what the sending and reception can llevar simultaneously.

Another point to stand out of this network, is that each node takes a time virtually zero (in the practice, lower of 1 µs) in reading and write in the plot that travels by the network. In fact, it says usually that the plot that goes through a node EtherCAT is going out to the following node before it finish to go in.

In figures, EtherCAT is so fast that it could freshen 178 axles, exchanging each one 64 bytes, in 1 ms. However, any current industrial controller is able to process this quantity of information in so little time. And of course, any another industrial network arrives to this level of performance.

Like this, in the market, the most advanced controllers will be able to control (no longer only exchange information with the same) some 32 axles in 1 ms, with what the neck of bottle, that in the past found in the buses of communications, now happens to the own controllers, allowing that these evolve of here to some years without having to concern by the system of communications.

The integration of the part of PLC, of ‘Motion Control', of vision, of robotics, of ‘Safety', of CNC, etc., no only brings achieve a homogeneity to the hour to develop but it allows a total integration between each one of the disciplines to implement allowing share variables, data and a synchronisation and absolute precision between each one of the elements of the machine.

At the same time, the software evolves creating true studies of development, from where can configure , no longer only the controller, if no each one of the distinct parts of the machine, from an interface common.

And this same software will have to be the attendant to guide us during all the process of development, assisting us and helping us in each one of the appearances of configuration and programming. But further to allow us program a concrete device, will be able to allow us and assist us in the creation of a complete machine.

A step even more there, will be the comprobación and the testeo of our development. The best current tools already allow to realise a simulation of all the disciplines implemented in the machine (sequence, control of movement, visualisation, etc.) and detect and correct errors before even to have the hardware. But no only this, but the design in 3D of the machine and the integration with other tools of regulation and processes, like Matlab will allow from the tool of software design all the machine and ensure, even before mounting it, that works.

Does not fit doubt that the future offers us controllers faster, powerful, with more memory, etc., but at the same time with useful functions, that do more reliable the machines and that allow a much faster development.