Are PLCs still relevant?

[horsefeather]

Well, I suppose I might as well step in it and display my ignorance.

I've been involved in flight simulators, spacecraft and automatic test equipment for 20+ years. The relevance here is real time control and monitoring of 1000+ sensors and effectors.

My recent endeavor is telemetering a hydraulic system with approx 20 sensor/effectors. The vendors keep trying to sell me PLCs and data loggers that appear to have been designed in 1980. Operator displays and I/O apparently require an additional laptop.

It seems to me that there should be a system that combines the Signal Conditioning Equipment (SCE) and the HMI (Human-Machine Interface) much like we do in simulators. I'm looking at the National Instruments product line because it provides graphical programming, simulation capabilities and an architecture that I am comfortable with.

The Allen Bradley and GE Fanuc lines I've seen seem to be bound by equipment designed when memory and processors were expensive.

What am I missing here?

 

[itsmoked]

I don't think the processor price has anything to do with it.

PLCs from day-one were abusively priced. They want absurd amounts of money for them.

One of the last ones I got a quote for was $13k for equipment that could not have cost more than about $600 to produce. A 22X mark up. They recognize that their industrial customers have deep pockets.

They also know that their customers are fairly gutless and are just trying to get other more complex solutions working to create their various products and a PLC is very convenient and proven. Go-the-freezer-get-the-box kind of solution.

Remember when the rule was "buy a PC from IBM"? Their stuff was slow and expensive but as a buyer you were protected because if things went pear shaped, "Hey! It's what everyone else buys!"

There are other systems you can use instead now.



Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[EricRICPDAS]

There are low cost PLCs now in the $100 range, which is quite relevant.

If you want a graphical user interface, you may want to step up from PLCs to PACs (Programmable Automation Controllers). See

http://en.wikipedia.org/wiki/Programmable_automation_controller

for a good discussion. They can do the control and human interface in one low power high temperature tolerant package. The WinCon,

http://www.icpdas-usa.com/products.php?PID=2971

is a good example. You can do the ladder programming as with a PLC in ISaGRAF, but also have a local user interface with touch screen or web interface with Indusoft.

Eric R

 

[PNachtwey]

Well, I suppose I might as well step in it and display my ignorance.

or our ignorance of knowing exactly what you want to do. In a earlier post I thought when you mentioned closed loop you were talking about closed loop control. You later posts hinted that the fluid flows in a closed loop so I am now confused.

PLCs are still relevent in that they don't run on Windoze or have hard drives that fail. You mentioned that your application required monitoring what happen when drilling. This is a nasty environment. Are the PCs and PC card you are going to use rated to run at 70 degrees C or in the cold? What about optical isolation? I know that PLC go through environmental testing that a Dell PC does not.

PLC manufacturers don't change standards every other year. I have PC cards with a ISA form factor that I can't use any more. One can still get cards for a PLC5 but they are expensive.

NI's labview works well as a data acquisition computer but PLCs work better as controllers. PLCs are not 'distracted' by memory management or I/O that isn't responding like does is. However, PLCs don't have the memory capacity of the PC. I would use the best attributes of both. From your previous post it looks like you have a oil well application. Time is money. I would go for what is reliable and safe above all else. A thousand dollars here or there is pocket change when the price of oil is $100 a barrel.

I think that you should consider using the labview with other field devices that can functions as controllers. The Labview can run on a PC in a nice environment and it can be connected to the working parts using Ethernet or some other field bus such as Profibus DP.

 

[horsefeather]

The closed loop referred to in my posting in the fluid power forum referred to a hydraulic closed loop circuit. There is a closed loop circuit for primary hydraulic power and an open loop circuit for the auxiliary circuits. I don't think they are relevant here. Different discipline - different questions.

I would use the best attributes of both

Kinda describes my frustration at available architectures. Of course this is based on only a week of research from on catalogs, brochures and vendor proposals.

From an electronic standpoint, the control would be considered an open loop servo with a human in the loop.

Environment concerns are valid: +120 deg F) as well as moisture (NEMA 4X) and fire hazard (Class 1 Div 2). My interest is in the architecture. Fiber optics are way too fragile in this application.

Sampling rates do not exceed 4Hz, so blazing speed is not a factor here. What I want is an architecture that supports modelling and display generation along with the SCE I/O. As stated earlier there are less than 20 devices in play, hardly worthy of a separate PLC.

There are several automotive PACs for <$4K that look appealing. The graphical programming language in the NI products are tipping the scales in thier favor.

Actually, the SCE is at the bottom of my list of worries, display generation, algorithms and portability (all software) will be my bottleneck.

Thanks for taking time to reply.

 

[itsmoked]

Regardless of the pricing issue they are still relevant.

There are several that are small, and you can "plug in" touch screen displays to for a nice, flexible, assembly line tough user interface. I was just looking at a system that was very impressive in that regard. I'll hunt it down if you're interested.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[controlnut]

You are kind of comparing apples and oranges..
ie - PLC Vs NI Data acquisition hardware
PLCs were designed with industry in mind, the reason you see designs that appear based in years past is primarily to provide backwards compatibility for the industrial market. Most factories don't want to redesign their control system every time a new widget gets invented. That said, PLCs are still evolving, and are still a viable control package. For more reasonable pricing and some PLC packages capable of what you described above, including embedded (windows) controllers , touch-screens, and HMI Software try

http://www.automationdirect.com

I have used their products on quite a few products.

Regards
Controlnut

 

[jraef]

IMHO, the issue at hand is fault tolerance. PLCs are made for factory floor environments and/or to be mounted on machinery; think vibration, heat, dust, cold, coolant, cutting fluids etc. One part can go down and you get diagnostic information on that specific part, not a "blue screen of death". Backplane and remote communications to I/O systems is optimized because they don't need to make info available for open source systems, so throughput time can be minimal and consistent. PLCs are also designed for machinery life cycles, i.e. if you install one now, the components for it will still be around for (hopefully) as long as your machine is relevant and useful. PCs are outdated in 1-2 years.

The HMIs that go with PLCs are expensive because they typically are designed around the same factory floor environment specs that a PLC would be. But there is nothing stopping you from running your GUI on a $150 laptop if you want to.

PLCs were never designed for serious closed loop control and direct HMI control. Those features were added mainly out of convenience. They paralleled DCS systems for that reason: DCS for the high-end big-picture process control stuff, PLCs for the discrete machine control. So called "PACs" have blurred the lines somewhat now, but PLCs are also getting closer to the PAC concept anyway.

 

[mikek10]

I'm not entirely sure of the specifics of your application, but for non-trivial servo control as well as sensor monitoring I would look into motion controllers.

Baldors Nextmove E100 is my current favourite and they also do HMIs.

 

[horsefeather]

Thank you all for the posts and discussion.

I go to Canada in a couple weeks to interview vendors. I'm resonably certain that I have the correct architecture in mind. It's just a matter of finding a vendor who can execute without proposing three computing platforms and multiple languages.

 

[ozmosis]

horsefeather
as long as your specification is clear, there is likely a solution out there. The "traditional" PLC format may not be right for your specific needs but there are huge variations on the theme of PLC's. For example, Siemens and AB both do a Motion Controller built into a PLC (or is it a PLC built into a motion controller?). I know that Siemens have a product called Simotion that might be up your street, not sure of the AB system. In your application it's about speed of data transfer and connectivity to devices.
If the specification is correct then you should get the right product if talking with the right company (and more importantly the RIGHT PEOPLE)

 

[controlsdude]

You know I have talked a few experts on using a PC versus PLC. I know A_ tried to sell chr____r on softlogics with a PC, greatest thing since swiss cheese. Later they kept getting the blue screen of death, but the PC kept running the process but you could not access the control application. I have heard this from other people, the kernel that the control app runs on will still run but it is impossible to access the PC to reset the problem. So you have to do a Microsoft reboot, which is not a good thing when its still controlling machines.

But in the plc side you can still get online with a faulted plc and reset the fault and also find where the problem is versus the blue screen of death.

 

[mikek10]

PCs have a place in control. They are cheap and easily replaced if there is a problem - so long as no non-standard components.

Its (relatively) simple to create complex (sequential) systems which has an inherent front end. Integration with reporting (eg Access) is simple. There is virtually limitless storage.

Backup software is cheap. Modem support cheap and simple. Etc etc. Communication via serial link, ethernet powerlink or fieldbus or similar is needed to communicate with servo amplifiers. Not too rugged unless you get an industrial rated PC which will defeat some of the benefits. The operating system can be an issue as mentioned before - on the other hand you can get non-skilled people to email or delete a file or what have you. With a PLC you need someone skilled - and remember they need a PC somewhere to communicate with it..

PLCs are the best for IO monitoring and are more reliable and faster. Not so easy for sequential operation but much better for non-sequential. They also start cheaply but get expensive if you want anything slightly off the wall.

Motion controllers fill a gap somewhere between the two.

To get a front end on the PLC or Motion Controller you need an HMI - for anything other than a basic display a PC and touchscreen is as cheap.


It really depends on your application what the best solution is.

 

[KitoOne]

if your into the VB you can consider like the SixNet, its a linux based PLC.

Elseif look into Beckhoff, its a very customizeable PLC but its not very heavy duty