PLC

I'm not an industry insider, but have had a look at the industrial control world. Here's my take:

A PLC is really a complete computer system designed for the control of industrial machines and processes. PLCs have I/O cards that they are designed to work with, the drivers are all integrated, and the operating system is very simple and reliable. PLCs will often run for years without maintenance or restarting. PLCs are programmed in several different programming paradigms, the most common being "ladder logic," which is a graphical programming methodology that does not require the programmer to know a software language. PLCs are also general purpose in that a given PLC can be used for a wide variety of applications.

By contrast, microcontrollers are more "do it yourself" systems which need to be designed for each application and programmed in a conventional programming language. They're generally more suited for embedded control.

PCs are starting to cut in on the PLC controller market. An industrial PC has many of the same characteristics of a PLC (standardized I/O with drivers etc..). Programming software is available that makes industrial PCs program much more like PLCs. The advantage(s) of a PC include: generally cheaper hardware because the components are produced in huge volumes for the consumer market, higher capability because the processors, disks, and memory are more advanced, and more flexible programming because you can always revert to "code" if you need to.

To my knowledge; PLC is Programable Logic Control using micro controller so; it is not just the micro contrllers. it includes the software of the program, the circutory logic and the micro controller. Again Micro Controller is the control element within a PLC system.

Expensive Tough reliable redundant controls with unlimited I/O and software capabilities use a limited production highly specialized reprogrammable PLC.

Small inexpensive disposable controls of limited life and specialized specs use mass produced microcontrollers that may or may not be reprogrammable.

The low cost, high volume, home PC falls somewhere between a PLC and microcontroller.

PLC is ready to use system and Microprocessor or Microcontroller requires engineering to make PLC. Visit some industry near your place and see, what PLC is doing there. Even gate keeper of the industry will tell you where to look for PLC.

What has been said to date is true but I believe a little history will give you a better perspective

PLCs started back in the 1970s and were designed as units that could be versatile, simple to use and install and be relatively cheap way of controlling various processes. What they consisted of was a microprocessor based Central Processing Unit of some description and a whole host of Input Output modules. The I/O modules were designed to interface with the real world and so could sense things like there being line voltage 110/240 volts, then switch these voltages and be able to cope with normal sorts of load currents of up to 10 or 20 amps. Some units had IO modules that were built in while others had IO modules that could be added on according to your requirements.

Back when these units were first developed memory was extremely expensive so you had a separate unit that was used to write the program which was then compiled into the appropriate machine code and loaded into the memory of the PLC. It may sound increadable today when memory cost a hundred dollars or so for a Gigabyte but back then it was not uncommon to see units with only 1KB or less of memory and in the early 1980 the best you could get had about 4KB of memory.

Back then when you were developing the controls for a specific process you would end up drawing a ladder diagram so to make the PLC easy to use that's how you programmed them. As time progressed Personal Computers came along and eventually applications were written so that you could use a PC to generate and compile the program for your PLC. Progress over the years has blurred the line between the PC that is used to program the PLC and the PLC itself. Even today though if you wish to control anything that requires you to switch things that rum on mains power then a PLC is probably the way to go because you can buy everything you need off the shelf.

You could of course do much of this with microcontrollers but if you wish to turn on something as simple as a 240V light bulb then you are going to have to do a hell of a lot of work building the circuitry to go with the microcontroller to do this. It's much cheaper and easier to use an off the shelf PLC.

Sorry for being so long winded but I think the history explains a lot of why things are the way they are.

In 1970 there were no Microcontrollers. Even PDP-11 computers were using CPU made of 74 series logic ICs and even DTL were used along with TTL. Digital developed LSI-11 much later. Only uC were SCMP by National Semiconductor and Perhaps Teletype paper tape and card punched data storage medium. PDP-11 mini were having 8" floppy and with small change in mains frequency, it could never read or write data properly. Phase lock loop or all new methods of data recovery came much later. I have used one 2708 EPROM some time in 1974-76 that needed +12V +5V -5V and +48V pulse to program. I also used memory chips that were having actual fuse inside the chip to be blown to write data bit by bit by fuming out material link. That was horible time. I think you are talking about State Machines or Moore Machines that were initial logic for all input and output sequencing to put many functions in sequence as now is done by codes in uC and uP. Even Facit calculators that were purely mechanical devices were used.

However technology sure was on the way. IBM did develop injet printers in 1974. They used to do page printing like we see data on CRT. They used to through ink bullets with charge and used to diflect the ink on spot else used to dump in waste basket to have blank spaces. They was no way to stop the ink flow. They used to have hammer based line printers. Teletype and other table top printers used to have 3D ball with X-Y-Z motion to have all characters on the ball for printing. Teletype were purely mechanical wheel and lever driven printers including paper punch and paper readers. I used to store code on then and load them. It was fun to write serial read write of data.

I made first transient data recorder using that 2708 EPROM with so many voltages. That was by silicon memory file. Today we have FLASH modules. I think we were not able to get 256x4 SRAM easily. In one system I was having magnetic rings for data storage and that was from ORTEC and was an MCA. IBM and Russian computers were based on those magnetic ring memories and used to occupy entire room and needed a power station to run.

You can't explain that world to the engineers of today. Most of us who are from an ancient era know the world differently. I was doing uC based designs from very first one and also have used State machine designs. They were very fast and the bit slice processors used them.

That is all History now. It is nano world today.

The 6502 microprocessor hit the market in 1975 and was used by Struthers Dunn in their PLC soon after.

You are correct in trying to explain things like core memory. It is a shame that much of this old technology was deliberately destroyed when it became obsolete. I remember deliberately trashing a great deal of very expensive equipment so it couldn't fall into the hands of third party maintenance companies.

The first computer that I got my hands on was an IBM-360 which if I remember was the first or close to the first to use integrated circuits. The actual machine I used to use is now in the Poser House Museum in Sydney. I think the Wireless Internet card that plugs into the PCMCIA slot in my laptop has more power in it than the IBM-360, its certainly got more memory. The thought of a Gigabyte of any type of storage was outrageous let alone having a couple of GB of RAM.

One thing you did learn though was to write efficient code, it makes me wonder how much surplus code there is in the programs that are circulating today.

Gee, I'm pretty sure I still have a 2K ROM board with 8 1702 EEPROMS on it around here somewhere!

I also have at least one 8K RAM board that used 1K static RAM chips (I forget the number offhand).

Both boards are for an S100 based computer (which I also still have) built on a Zilog Z80 8-bit μP chip. The same cpu that now runs my son's TI-84 calculator (how's that for a long-lived design in the digital era).

Z-80 uP came in 1978 or near to that. It was partially a copy of Intel 8080. Intel also had a 4-bit version. There were some Rockwell chips used in programmable calculators but were never sold for uC developers otherwise. Calculator chips with math functions made their own end products.

Most Interesting development was 12-bit Microprocessor developed by Intersil, that never came to the market. They failed miserably.

Another important development of the old timer was development of DAC at Teledyne Philbrick. These DACs were in cubic modules and I removed their packing to find details. They had PNP transistors as current switches, current mirror to form voltages as accurate as 16-bit and real fast that they could be used to scan RADAR screen displays in real time. I think, it was Bob who is now at National Semiconductor, did that development.

Today's ADC and DAC are so common in PLCs but their developmental history is very interesting one.

Earlier computers were switching programs as forground - background modes. Then came concorrent software plans. In 1978 I already designed multi-tasking system for nuclear MCA using Intel 8085 uC and fast state machines using PROM. It worked with about 20000 talk switching per second and is still in use. As this 8085 was only uC built radiation hardended, it never failed for the last 28 years. Great development.

Apple computers amd stolen idea as IBM-PC from Apple came a bit later.

People of that developmental era know things deeper in design as they did everything by hand, wired their own computers and coded each byte with lot care. There wasn't enough memory in the system, and program storage on paper tapes was the only way to reload the code in RAM. Earlier Space Missions also used those 256x4 Static RAM as nothing else was there. I think they used only 32kByte memory for image data.

I agree that Z-80 as Sinclaire Hobby computers that came for games made a lot of difference. They remain popular for many years and reached home. Code in them was in small EPROM.

Texas also built one 100-pim golden uC 16-bit version 9900 series and was used by companies like Lecroy and others in their oscilloscope for data handling.

I also remember that there were transient analog signal recorders with dynamic RAM with only single input and output pin and data was always rotated inside using continuous clock. They were wonderful devices with simple technology. I could catch microsecond explosion shock waveform using those transienct recorders.

I think earlier industrial machines were Numeric controlled and then came CNC machines. PLCs were more like dedicated logic.

My earliest encounter with PLC was in 1988 and by then 200 types were already in the market with Ladder diagram language and direct code language.

Technical manual of IBM-PC was top secret and only for its employees. It was sold for US$100. IBM mainframe were using 3"x6" paper punched card for storing codes. Their data storage tapes used to to spoolers rotating under vacuum to gain speed. Perhaps a pen drive today has greater memory than those man-height magnetic tape spoolers.

I used those wirewrap machines to quickly interconnect the logic and make control system of very large size in few days. Bishop Graphics also did their good contribution in making PCB stensils and color tapes.

If you have seen electronics of very earlier times - then there were only valves as singnal aplifier and power drivers, many plates as capacitors, and heater coils as resistors and inductors. No real components of todays world was seen in the instruments. They were parts of the world war II. I have seen these in 1970 while in my MS Physics education. They worked well. They were in 4U height 19" rack modules size to make a wireless transceiver. I think microphone and speaker were samething and were made of carbon powder in a casing.

You get PLC for US$50 today. You can control your entire house and your garden out of that easily. PLC is to be used and not to be talked about. It is simpler than computer programming device. See how simple it is

Input # ----------|gate logic AND OR etc|----------(Output #)

Have fun and use them.

I remember working on the Honeywell H-316 computer back in the late 70s. It used core memory that was the most advanced available at the time. an 8Kbyte unit was about 20cm x 10cm and was about 2cm thick. Core memory doesn't loose it contents when powered down so I had a module with diagnostics in it that I plugged into machine that I was working on. When you got a new machine the memory was all set to zero or null and you had to load what they called the Key In Loader (a simple program to read from the paper tape reader, card reader, magnetic tape or dick drive) bit by bit into the first 20 memory locations. Everything back then was in octal and when the μP guys started using hex there was a saying;

REAL MEN COUNT TO EIGHT

For its time the H-316 was an incredible machine, the CPU was made of surface mount DTL (Diode Transistor Logic) chips spread over 11 circuit boards. If I remember correctly the whole thing was built using about 3 chips a quad NAN, a quad NOR, a quad XOR plus a handful of germanium diodes and the backplane was all wire wrapped. The only things you needed to carry around in your tool kit was a logic probe, soldering iron and a handful of chips and diodes. Whenever you needed to repair the thing you would get it to execute each op code until you found the offending code then you had this incredible manual that had a flow chart showing how each op code worked and next to the flow chart was the circuitry that actually performed the operation. It was a simple task using a logic probe to trace the signal through to find and then replace the offending chip. I think I still have a copy of the manual, I kept it because I thought the concept was so brilliant.

The H-316 was reworked to use Transistor Transistor Logic TTL and ended up the basis for the PR1ME 50 series machines. Later on some of the processors used Current Injected Logic (I²L IIL etc.) basically OP amps used as logic gates.

One of the techniques I saw used for multi tasking machines was to have multiple register sets, one for each of the processes that needed to be executed. A process exchange could be executed between clock cycles by switching to a different register set, something I believe μPs have yet to achieve, I could be wrong though.

Things have certainly changed, if the automotive industry had kept pace with the computer industry we would all be driving Rolls Royces that were the size of a match box on the outside, could fit an Olympic swimming pool on the inside, was able to drive to the moon at the speed of light using a AAA battery for power and cost $0.99.

H-316 yes. We had one called Honeywell Bull and we also had one called BASUM-6. You rightly said that people at that time looked for reliability that could go for decades. However, technology moved much faster.

Industrial electronics like PLCs were great idea to use the programming capability and all industrial reliable parts. Many people do not know the difference when we talk about reliability. Let us see, how reliable a relay switching will be in long run. If I have 1-million operation rated electromechanical relay, then how long it may last?

If we switch relay each second and there are about 31,536,000 seconds in an year so one million operation switch will last only less than 12 days. However, PLCs supposed to run for years. Hence, if it has 32 million switching life then may last for an year. I looked for such switches and found that COTO Japan www.cotorelay.com was making 1 billion contact rated relays, while most of the industrial relays are 1/1000 factor down in reliability and may last only few days.

There were mercury weted relay to avoid contact bouce. They are now banned as they have poisonous Hg in them. Now MOSFETS and optically isolated zero cross relays are replacing the old style electromechanical relays in PLCs.

There will be need for greater high voltage relays in PLCs that can go up to few thousand voltage, switch faster, and drive special servo controls with all shorts of high grade control mechanisms. Today's world is high ly demanding type and technology changes each year and not in decades. You can't avoid superior and cheaper technology taking over the existing one.

I now use high voltage switches up to 30000V that switch in just 25ns and can drive 60A into load and these sit in small casing in your palm. Look at the power they can switch and remain cool. I can make a train to run from these.

Is industrial PLC technology saturating in design now? I will say now. Old stuffs are going out and new ideas are poring in. Industry can slow down but will never be wiped out. It can also have different Avatars. Then there were Fuzzy and Expert PLCs using Artificial Intelligence. I am sure there will be human brain equivalent or even superior brain PLCs.

I started from an old era and moved to beyond todays technology with creativity. Horizon of saturation was never near in the last 30 years of my journey in this field. I think it is fun, pleasure, money, life in working in this zone of knowledge. PLCs will get new name soon. Perhaps I will like to have some ALCs or automatic logic controllers like human brain activated and conscious machine controllers. I have invented this new name for many people in the future to look for their future in it, far greater than PLCs and computers controls. Have fun!

I did some very extensive work both with and on the Struthers Dunn 1001, 3001 and 4001 PLC in the late 70s and early 80s so when this thread started I had a look for what these devices wore now worth. To my surprise a fully optioned Director 4001 PLC CPU is worth around US$5,000.00 which is not a lot less than what they were worth new. I am surprised that there are still any around after a quarter of a century but the price that they are fetching tends to suggest that they are still in use and in demand. This could also mean that there is nobody left in the industry that knows how these units worked and the only option when a unit fails is to replace it with an identical unit rather than upgrading the technology.

The fact that these units are still in existence after 25 years of use proves your point on reliability. Try and get that level of reliability from mechanical relays!

Why the undefined <...we...> does any particular thing is not the sort of question that a global anonymous forum can answer. Considerable research would be necessary, and any invitation to do so is declined out of disinterest.