PC vs. PLC control


Thread Starter

Stephen Wright

Will someone share some unbiased thoughts on the relative merits of PC versus PLC control? I shall install a new conveyor system, connecting
several machines. The bright idea I had was, since I needed some operator interface screen, might I upgrade the OI to a computer, use it to control the process and eliminate a PLC. PC vendors think this is a great idea, PLC vendors warn me about MTBF. This is a relativly small system, about 15 motors and 100 discrete I/O points. The environment is rather dirty, and
operator abuse of operator interfaces has occured in the past.
When Micro$oft Windows does it's usual thing (crash) while your system is running, you will figure out which was the better decision.


Jocko Harmet

I have worked with both in the past, and my experience has been that the PC makes the best operator interface you can get, but by the time you make it into a PLC you have a great HMI with marginal io performance.
Some manufacturers make Ethernet IO that is designed to be connected to a pc running there hmi software - and for small applications this works well.
In my opinion, you would save a lot of headaches with a combination of both. PLC's on the plant floor and a PC running HMI software in the control room.
something that should be considered is that the PC is designed around information processing and graphics - in the office. How often does a PC crash? what state is it left in? are the IO on or off? What has happend to your process in the mean time?
A PLC is designed to handle lots of IO, and be rugged enough to withstand the factory floor environment. How often do PLC's crash? almost never. when it does crash, all IO are turned off.

If you are looking to the PC as a cost savings, in most cases a properly engineered system using both a PC as an HMI and PLC's for the IO the savings going with a PC is minimal. Watch out for hidden cost in the PC system, drivers, custom software, etc.
If you have looked at a PC software package, make sure that you can do your application with whats there.

Jocko Harmet
[email protected]
623-580-1101 fax

Jason Halligan

PLC/PC we use several pc operated machines in non vital/safety applications, if a non vital machine stops due to power failure etc, who cares it may take a few minuites to get up again, bugger.

PLC/PC on essential services like Amonnia Plants and Standby Generators the PLC has a major advantage of up time in millseconds.

You need to grade/quantify your actual requirements to come up with the correct solution.

Jason Halligan,
[email protected]

Anthony Kerstens

The merits of PC versus PLC controls are irrelevant.

- These machines being connected are controlled by what?
- For the small requirements, could one of the processors in any of these machines be extended to a remote I/O rack to control new stuff?
- Similarily, do any of these machines have an HMI already built in that could be extended?
- Do the people maintaining these machines understand PC controls?
- Are PC controls compatable with your spare parts inventory?

Start by writing out your current investments and evaluate all the costs of going with any variety of scenarios. List things out, add up the costs including parts, software, and training.

It may work out that PC control makes sense in your situation, or it may work out to be a frustrating exercise. If you have the least little bit of doubt, go with a PLC and save the cool stuff for a later project.

Anthony Kerstens P.Eng.

Terry Miller

Stephen, the issue as it relates to moving from PLC to PC control (Open control) is what do you want to accomplish beyond simple machine control. My company is a producer of Open Control software but I always tell people to look at the architecture needed to accomplish your goals. If all you need is machine control then PLC's are the choice; they are the best, most reliable
control invention ever to hit the factory floor.

If you want/need to take advantage of PC power for data acquisition, HMI, alarms, trends, etc., in addition to control then Open Control will probably serve you better. Only by adding or consolidating these functions with control will you be able to take advantage of the power of PC's plus create a more cost effective solution. If you need connectivity to other legacy
(proprietary) systems either networked or web-based then you also will be better served with open control.

As far as hardware reliability, most of that is smoke and mirrors. Hardware form factor with PLC's is relatively fixed. With PC's you have options to add control to flat panels, or simply to tailor the footprint to the size and scope of your system-single board computer, industrial PC, shoebox, etc. There are many more platforms serving the open control arena than the PLC one, and all have extremely reliable platforms available.

I hope this helps.

Terry Miller
FlashPoint Automation,LLC
[email protected]

Bob Peterson

> Will someone share some unbiased thoughts on the relative merits of PC
> versus PLC control? I shall install a new conveyor system, connecting
> several machines. The bright idea I had was, since I needed some operator
> interface screen, might I upgrade the OI to a computer, use it to control
> the process and eliminate a PLC. PC vendors think this is a great idea, PLC
> vendors warn me about MTBF. This is a relativly small system, about 15
> motors and 100 discrete I/O points. The environment is rather dirty, and
> operator abuse of operator interfaces has occured in the past.

Personal opinion only.

I presume you are refering to some form of softlogic PLC when you talk about PC control. For a 100 point system, you are unlikely to find a software package and reliable PC hardware that can emulate a PLC for what a real PLC costs. Go with a real PLC.

As to putting the OI and the PLC ont he same hardware, this is of dubious benefit in such a small system cost wise since you can buy a real OI for probably less then the software you might buy for the PC.

Bob Peterson
+ professional programmer loves it
+ more (and open?) software/hardware choices
+ sounds more high-tech
- more integration job/responsibility for developer
- end-user maintenance engineer may not feel comfortable to handle
- costs more money for robustness
- harddisk may fail
- boot-up takes longer time
- spare part may be a big problem after few years

+ robust with minimum cost
+ except high level programmer, everybody likes it
+ no integration job/responsibility on controller level
+ guaranteed spare part supply
- locked to the vendor
- sounds old fasioned

Michal Casterline

There are indeed a number of companies that make PCs that will fit in an MMI form factor.

With PC-based control you will likely have to pay for a software license, but in many situations the installed cost is still competitive.

PCs are blindingly fast and very flexible. Still, some prefer the comfort provided by the limitations of a PLC.

If you do choose to proceed with a PC, consider a solid-state hard drive. Rotating media is the Achilles heel of a PC in industrial applications. The environment you describe is no more a problem for the electronics in a PC than a PLC.

Michal Casterline

Ranjan Acharya

This debate again ...

For a small non-critical system I do not see the difference between PLC or PC based control. We have set up systems with the PC version of RSLogix and an associated RSView HMI "all in one" on reliable hardware under plain NT (no other applications!) and it runs just fine. Ruggedised PCs I installed years and years ago for HMIs are running just fine too -- not a statistically accurate survey, but I think with the right hardware you are fine. Fan-less, ROM-based solutions with solid state drives are always good
things to consider for nasty environments -- that is what your PLC is, it just has a better real-time OS than Windows NT provides.

That being said, I personally prefer PLCs for most control applications. I still do not trust Windows NT (or any flavours thereof) for long-term
real-time control and our customer base has not embraced Linux yet (nor has Linux been embraced by the major control system vendors [Schneider-Modicon, TI-Siemens, Allen-Bradley/Rockwell ...) with a commercially available version of RSLogix for Linux and so on). I am not trying to spark a Linux versus Windows debate here, that horse has been flogged to death. There will be lots of juicy replies flaunting Linux.

You said that you were concerned about operator abuse of HMIs. That would steer me in the direction on a traditional solution with a PLC and HMI, not a combined PLC HMI on a PC. There are small PLCs that will do your job for less than CDN$1500 (programming software probably extra, it is a long time since I looked into pricing of small PLCs).

Of course Linux is absolutely free. You just plug it in. :)

David McGilvray

Don't take my word for it, as I am biased, just look around you. Mission Critical (ie down time = very high cost) applications are running on
PC/server environments as the norm in virtually every area outside of process control, including banks, brokerages, internet sites, etc. etc. etc.
While I can't necessarily vouch for the product you are considering, the fact that it is PC based can be an advantage as you can choose the platform
(hardware, OS, etc) that suits your requirements (reliability, supportability, etc) in stark contrast to most proprietary product vendors who offer you any colour so long as you want black.

David McGilvray, P. Eng.
M&R Automation

Andrew Piereder

I think you need to ask yourself what the impact of your controller's faulting would be. I have often been in situations where the need for a
reset on a comm link every four or five months was considered completely unacceptable. PCs generally crash somewhat more often than that. If you can live with that, then PC control is probably a cost effective choice, if not, then back to PLCs and OIs ;-)

As for the environmental situation, good industrial cabinets exist as do rugged industrials industrial PCs, however you do end up paying more and lose some of the economic incentive to choose the PC over the PLC in the
first place.

Andy Piereder
Pinnacle IDC

Curt Wuollet

I think you have come to the wrong place for unbiased thoughts :^) But it depends entirely on what you have to do and the tools you
use. Can you whip up something in VB and run your conveyor reliably? I would say no. Some would argue with this of course. Could you use
something like steeplechase? Probably, but cost and complexity would be overkill. SoftPLC would be an option I would be comfortable with but still would be pretty pricey for this application. We have done a sort conveyor with a PLC and a PC running Linux very successfully
with the PLC doing the logic and control and the Linux PC handling several barcode readers, communications and display. We did it that way because of time constraints. The respective functions were already developed or quick to develop. Since then I have developed the tools to do it entirely with a Linux PC very inexpensively,
quickly and reliably. That is how I would do it now, because I have the tools. I will gladly share them by the way but there would be a learning curve and you would need to know C and fab some boards. The LinuxPLC would be very nearly ideal for this type of "all in one"
application, but it's not ready for prime time yet. It would do what Steeplechase, SoftPLC etal. do. Without the software cost. This would make it cost competitive with hardware PLC's.

There is a fairly recent class of what amounts to a display with IO added. These might be quite competitive here also. They were developed with the low end market in mind.

If you can't do it that way, bear in mind that some PLC vendors have PLC's specialized for this task. GE, I know has them. If they fit well that would probably be the least cost. The display
part for that would be Horner OIU's which I have not found to be very reliable or cost effective YMMV. I have used Matrix Orbital displays and they are very inexpensive (in automation terms) but they are much easier to use with a PC solution than PLC's as the serial comm facilities on PLC's tend to be tedious unless "basic" or "ascii" modules ($$) are added. Optimation has inexpensive terminals that interface with popular PLC's. Because of the pricing structure and
programming environment in automation land "low end" projects are quite frequently best done in what you are most familiar with as it is painful and often not cost effective to jump into something new for a "one shot".

I just thought I'd respond in a relatively unbiased way before you get the "whatever I sell is the best for you" responeses. I sell nothing.

Hope this helps



Donald Pittendrigh

Hi All

> It may work out that PC control makes sense in your situation,
> or it may work out to be a frustrating exercise. If you have
> the least little bit of doubt, go with a PLC and save the cool
> stuff for a later project.

Sorry to have to interject but I detect a derogatory tone in the use of "cool". I have spent 20 years in the PLC game and was difficult
to convince, but, when I think of Soft PLC today I don't think cool I think Fast, Cheap, Ease of use, simple to configure hardware components, I think of the smirks on the faces of all the "proprietary vendors" I have dealt with over the years when they said, "you want to connect what? to our PLC" and then handed over the bill. I was convinced by the Think and Do PLC range and I think they may very well be regarded as cool but that's hardly their prominent feature

Now I only have to convince my clients that I am right!!

Donald Pittendrigh

Anthony Kerstens

Well, I'm sorry you sense a derogatory tone.

I seriously think it is cool.

Anthony Kerstens P.Eng.

Steve Myres, PE

It sounds like in your application, one of the prime criteria may be the physical/mechanical ruggedness of the OI / PC, in addition to any concerns about the control reliability of the platform you choose. This makes PC control even less attractive for you than for the average job, if it is your intention to use an office PC. If you use an industrial PC, there probably won't be any cost savings over a PLC, and you might even see a net negative.


Jake Brodsky

As you have seen, there are some strong feelings on this subject. I think most of the replies you've received so far have held those feelings in check very well.

This is very much an application and a perception issue. A very similar issue was covered on sci.engineering.control news group recently. Check the archives on google.com for details.

Without getting in to OS wars, software cults, or other such social trappings, allow me to relate my experience with strictly the hardware side of things: Can PC hardware be made reliable? Yes. Can PC networks be made reliable? Yes. Can a limited set of software on a PC be tested to prove reasonable reliability? Yes. Would a system tested to that degree be less expensive than a PLC and PC combination? NO. In fact, it might even be more expensive.

The reason PC hardware is generally regarded as less expensive than a PLC has much more to do with their cheesier construction and poorly tested software than it does with anything else. Once you have upgraded a PC to the point where it will survive an industrial environment, and the software to where you'd feel confident enough to trust it to run a dangerous industrial process, you'll have re-invented something very similar to a PC/PLC configuration and the difference in cost will be negligible.

So, if your application is critical, you know what to do. If you can survive down-time, you have inventory and staff to restore a PC to full operation, and you can keep your operations staff from loading any other software on that platform, a PC may not be a bad fit. The rest of us wish we could be that lucky.

Jake Brodsky
[email protected]

Cesare A. Perani

We use PC for direct control since 1992.
Our application are on-board diecasting machine.
PC direcly control the iniection fase (the most critical 50 microsec.).
In one software we have:
- hmi
- supervision
- data acquisition
- networking
- SoftPlc

Actually as software we use XPOLYPLUS and for the harware, NEXOR I/O, that is expressed designed for SoftPlc and SoftLogic application.
(nexor i/o has some features like autoreset of output, if the communication with the PC is lost).
As computer we use only industrial computer with disk on chip.
The performance are superior to any plc.
We can have interrupt procedures called by encoder.
Only one software environment for supervision and control.
Less costs.

[email protected]
PLC is much reliable than PC .in PLC the data mangement has done excelent than the where will develop in PC.
My strong recommendation is to go for PLC for these industrial and complex logic system application.
Asfaras my knowledge goes Allen-Bradley PLCs are very reliable
I have programmed and installed many PLC's in my 25+ years. I have PLC systems that have been powered-up and running for 10 years strait. If you can find a PC system that can run for one month strait through, I'd love to know about it.
Also, you must always consider the safety issues involved.