PLC + SCADA vs. WinPLC system

J

Thread Starter

Jeff

We are in the middle of design for one of the most complex fish hatcheries ever to be built in North America. Our engineering consultant is recommending standard PLC and SCADA alarm and control system but one of our staff is really pushing a WinPLC based system. I can find little if any info on the implementation of such a system. Are WinPLC systems commonly used in general industry? Or is this still an emerging field?
 
I suspect this is what the OP is referring to:

http://www.hosteng.com/HW-Products/WinPLC/WinPLC.htm

The WinPLC is a diskless runtime control platform that gives you all of the best features of both traditional PLCs and Windows capable PCs. The WinPLC allows us to take full advantage of all the recent advances in software technologies while still maintaining the ruggedness of the PLC type hardware. The WinPLC was the first platform to provide an industry standard operating system in a PLC form factor.

Really? Not. Deception and marketing hype.

Here's a reality check. Our firm remotely administers numerous Win CE devices via VPN and the internet. We have had to install IP addressable relays (brand name webrelay) in the power circuits of these devices in order to power cycle them because the O/S locks up so often. Record this URL,
http://www.micro-decisions.com/rio/webrelay.htm
if your associate is devious enough to win the argument. You'll be able to restart the PLC remotely from your desktop whenever it locks up. Hopefully the fish pond won't drain each time it happens.

Win CE might be fine for playthings in an economy of superfluous wealth. But serious control? Never. Run from Win CE. It is a very, very bad practical joke. It's the kind of thing one wishes on his enemy.

It doesn't matter what the form factor of a controller is. Serial murderers are packaged with charming, engaging personalities. But they are still serial murderers. A box that looks like a PLC but runs any version of Windows is NOT a PLC, never will be, and is a disaster in the waiting. It is waiting to murder your process.

I'm not sure what a fish hatchery controls, but I can well imagine sluice gates, drains, feed water, liquid feeding systems, level controllers and what not routinely going into fail safe mode when the CE box locks up.

Your engineering consultant is on track and is worth what you're paying him for. Your associate is a fool to recommend control through CE.

Any brand name PLC is built to be robust. Win CE isn't. It's fine to put your SCADA on Windows, it isn't controlling the process. Leave process control to REAL controllers, not to a PLC wannabee like CE.

David
 
C

Curt Wuollet

Having your control system blue screen on a weekend could really begin to stink.

Regards
cww
 
Dear Sir,

This question has been raised a number of times and answered a number of times, which system should be adopted?

I think almost all PLC/DCS systems are good to perform the desired operations. The only criteria you must use are:

-> Price of the system which suits you
-> The system you can easily handle
-> Last but not least, the system you can get help and assistance from the manufacturer the most easily

Regards,

A PLC Programmmer
 
Hello Jeff;

Most large PLC manufacturers are now offering their own version of "WinPLC" systems. Here are a few examples:

From Siemens, the WinAC platform, either purely software or slot-PLC: https://www.automation.siemens.com/simatic/pcbased/html_76/produkte/pc-based/based.htm

From Rockwell, a purely software controller:
http://www.ab.com/programmablecontrol/pac/softlogix/index.html

From Schneider, a slotted co-processor application:
http://www.schneider.co.uk/internet/pws/pws.nsf/luAllByID/10B1ED8DD933B84A8025710A005514BE?open

Success of these controllers in the market has been mitigated, due to control engineers and technicians being wary of PCs and Windows software stability issues (have a look in the Thermal Overload section in this forum, for threads on PC reliability "that wouldn't die").

PLC hardware and software is designed to run 24/7/365 with no mouse/keyboard/screen interface with a user, only with process data. Basically same chips as PCs, but software implementation makes cyclic processing of the program efficient and robust. There are Linux-based applications being developped (look at opensourceforge.com, the puffin project, for example) but I have not yet seen one working in an actual industrial project (Curt will probably correct me on that!)
;-)

I have see a few applications of the WinAC slot-PLC running well, especially when external power is provided to the PCI slot-cards, so that the PLC application will still run even if the PC crashes or has to reboot.

Notice that you are only swapping one hardware platform for another. You will still need the programming software from the manufacturer, and since you have no access to direct I/O cards you will need to implement network distributed I/Os (Profibus/DP, ControlNet/DeviceNet or Ethernet I/O).

Hope this helps,
Daniel Chartier
 
N

Nathan Boeger

Do you like sporks? I don't - but I might use one to eat a small pot pie. If I had to down a pot of soup broth and a T-bone, more useful equipment might be a spoon, fork, and knife.

I assume you're referring to Automation Direct's WinPLC. I've never used it, but get the idea based on their web site. It is a 40 or 100mhz CPU placed inside a PLC slot running Windows CE - like my cell phone. I'm a strong proponent of Automation Direct. This device looks very capable in terms of bang for your buck - particularly with small applications that require PID loops or complex math - or if you have a very specific application that you want to custom code in VB or C++ that can't run on a computer (arghh??). For the most complex fish hatchery in North America this makes absolutely no sense.

To the best of my understanding those types of systems are uncommon and unlikely to ever really catch on, just like softPLCs running on PCs. Control system experts are mostly diehard PLC bubbas. The convergence is taking the form of more sophisticated next generation PLCs. General PCs can't be touched in the SCADA alarm and control arena. It makes sense to go in the direction of the industry standard, for future maintenance and upgradability, if nothing else.

Keep us informed on your progress.

----
Nathan Boeger
Design Simplicity Cures Engineered Complexity
http://demo.inductiveautomation.com
 
M

Michael Batchelor

They do work, but quite honestly we removed the one we had in an installation and replaced it with a straight Automation Direct DL(something_or_other_that_escapes_me_now) processor.

Michael Batchelor
www.IndustrialInformatics.com
 
I've spent 10 years of my life working on Control & Automation, and never heard about that system. I'm not saying that is not a good system, but definitely I will prefer a PLC based system.

I think that you could give us more data about your process, and then we could help you to find a good platform.
 
W
Well, if you're using WinPLC to be the generic, as in Programmable Automation Controller, or PAC, yes, this is commonly used in industry. Examples include Rockwell's Contrologix, the actual AutomationDirect WinPLC itself, Advantech's line of PACs, National Instruments' line of PACs, Opto22's line of PACs, and many more.

Since fish only die once, I'd recommend being conservative on your design. If your engineering consultant is recommending a design, you might very sincerely consider going with it... after all, what the heck are you paying them for?

Walt Boyes
Editor in Chief
Control magazine
www.controlglobal.com
blog:Sound OFF!! http://www.controlglobal.com/soundoff
_________________

Putman Media Inc.
555 W. Pierce Rd. Suite 301
Itasca, IL 60143
630-467-1301 x368
[email protected]
 
B

Bruce Durdle

It's not very fast either - only 40 or 100 millihertz! (But it does have 4 megabytes of memory).
 
M

Michael Griffin

You have by this time heard a number of opinions on the WinPLC. To answer your questions more directly though.

Q: "Are WinPLC systems commonly used in general industry?"
Answer: No. This is a niche product sold by a minor vendor.

Q: "is this still an emerging field?"
Answer: No. The market shows no sign at this time of changing to this type of product.

The "WinPLC" is just the brand name for Think & Do's software when embedded into a special version of Automation Direct (AD) CPU. Think & Do is a company that has been selling "soft logic" software (emulates a PLC on a PC) for a good many years now. Their particular niche is that you program it in flow charts instead of normal ladder logic. The WinPLC simply embeds that software into a board that replaces a normal AD CPU (DL205 series).

Here's a link to the actual product.
http://web1.automationdirect.com/adc/Shopping/Catalog/PC-based_Control/WinPLC_Products/H2-WPLC3-EN

I have some experience with the normal AD PLCs, and have no complaints about them. They are the marketing name for Koyo, which is a Japanese company who has long made hardware sold under the TI, GE, or Siemens names. I haven't anything to say against Automation Direct (Koyo), but the WinPLC is a specialty product, and not part of their main product line.

The Think & Do software is fairly rare, and I've never laid eyes on an actual installation in either a WinPLC or in a normal PC form. Someone who had used their software some years ago though told me that although the flow chart programming is easy as a general concept, it is actually difficult in practice to do many normal PLC type operations with it. That is, you end up with a long and convoluted flow chart to do something which is very simple in ladder logic. I would expect to see higher programming costs unless you are dealing with someone who has a lot of experience using the Think & Do software. In addition, while a typical industrial electrician can read and understand normal ladder logic, they are extremely unlikely to have even heard of Think & Do flow chart programs, let alone have any experience with them.

Another thing about this, is that it is really just another (unusual) PLC CPU. It isn't a "base" to a solution any more than any other PLC CPU would be. You will still need the rest of the PLC rack, a SCADA, and a control network. What you should be talking to your engineering consultant about is PLC and SCADA brands based on the features you need, availability of spare parts,
cost, etc. Something that might matter more to your application than it would to a typical factory would be the humidity and temperature tolerance of the hardware. I assume that a control network (to sensors and actuators) will be an important part of the system. You should be trying to use one which is supported by the greatest number of relevant vendors possible. The control
network can be the most difficult part of the system to replace in future.

One of the previous replies mentioned problems with Windows CE locking up regularly in WinPLC installations. While I can't claim any first hand
experience with WinPLC, this problem with Windows CE is not uncommon in other applications. The main market for WIndows CE is in mobile phones (where they are a minor player in the "smart phone" niche) and it is notorious for this problem there.
 
M

Michael Griffin

In reply to Bruce Durdle: I believe you meant 40 to 100 millisecond scan. 40 to 100 "millihertz" would be 25 to 10 seconds scan (hertz is scans/second, so 1 millihertz would be 1000 seconds). I'm sure it's not that slow.

As for amount of memory, that is a poor guide to how much it can do. We don't know how much memory it needs to do anything, and that is something that varies drastically between different PLC designs.
 
C

Curt Wuollet

For the record, the tiny C precursor to the MAT/plc I wrote was successfully applied in an actual industrial setting and worked reliably for years until the company went under. But there is
really no way of knowing what people are doing with the project, it's theirs to keep so they don't have to register or even say thanks.

But, more importantly, I wanted to add that some very large, very serious, printing press builders have gone past PLCs in favor of networked high capability racks running Linux, RTLinux, and a hard RTOS. And the post press equipment went from
Mitsubishi PLCs to Beckhoff embedded computers with all networked I/O. They claim it's more reliable. The investment at risk is staggering so I'm pretty sure they have done their homework. This is all German equipment. The networks are all important and PLCs simply don't do networking on this scale. Just a datapoint of the latest high value systems I've seen. I think it's the start of the post PLC era. The tower of Babel collapsing
on itself in favor of stuff that _does_ work together.

Regards
cww
 
W

William Sturm

I think Bruce was referring to the fact that they use mHz on their spec page instead of MHz for the processor speed. :>
 
N

Nathan Boeger

Cumon Walt, he's not looking for a broad stroked BS answer about PACs, he's asking about a specific product line for a specific application that was recommended by his consultant. The product is for really small applications (that's being polite). His application obviously doesn't fit that mold. Besides, a Windows CE based control systems hardly fits into the same category as a Control Logix PLC.

OP was given a poor recommendation by that engineer. That's further evident by the sheer number of experts that shot down the idea so quickly.

----
Nathan Boeger
http://www.inductiveautomation.com
Total SCADA Freedom
 
Jeff,
Like most others I wouldn't like to rely on a PC based system. A PC is fine for display if it crashes the PLC caries on. We have used an inexpensive product from Opto22 on several recent projects. The software is free and what's really nice has no licence to worry about, will run on any old PC. Programmed in flow chart which takes a bit of getting used to after ladder but seems to be easy for a Non ladder engineer. check it out at opto22.com

Roy
 
M

Michael Griffin

In reply to Bruce Durdle: Sorry. I guess I didn't get the joke at first. I was looking at the Automation Direct specs, and they had the correct units (MHz rather than mHz), so I didn't notice this problem.

I do notice though that neither site (Automation Direct nor Host Engineering) have any real specs as to how fast the WinPLC with the Think & Do software actually is. Both sites tell us that the microprocessor runs at 40 or 100 MHz. However, since we don't know how fast the software runs we really have no idea how fast or slow the system actually is in terms of scans per second, boolean operations per second, or any other benchmark that is actually useful. The Flash/RAM size also doesn't tell us much either, since it tells us nothing about how large of a program we can actually load.

Another point that I noticed when reading the fine print is that only certain I/O modules are supported by the WinPLC. There also seems to be a problem with some revision levels of WinPLC not supporting some revision levels of I/O modules that are in the supported list. In other words, know your application needs in detail and read the fine print before using this product.

According to Host Engineering, you can buy a "bare" WinPLC without the Think & Do software. In this case you can buy a development kit to write your own software in C/C++. I can't imagine wanting to do this myself, but this question has come up occasionally.
 
M

Michael Griffin

In reply to Curt Wuollet: I have experience with a couple of PC based systems in a different application area.

#1 - With the systems from one supplier, they were well engineered and didn't use any hardware or software that was prone to failure. The PCs were diskless, fanless and used a QNX OS. Everything was designed for accessibility and had proper attention to cooling. The machines were very highly regarded by anyone who had to use or maintain them.

This company also had an earlier generation of product that used PLCs (which we also had). The PC based systems though were at least reliable as any PLC (and the software was far superior to the PLC based versions).

#2 - Another set of equipment from a competitor was also purchased which was somewhat more problematic. These were also PC based. However there were fan failures (followed by overheating), hard drive failures (the machines inherently ran under continuous vibration), and software problems (VB and MS-Windows based).

The difference between the two is one system was engineered for reliability, while the other wasn't. I doubt there was any significant difference in cost between the two though. I suspect that when most people think about "PC based systems" they think about ones like the second system, rather than the first.

As I have mentioned before, we had a discussion on this list back in 1996 where Dick Morley (considered by most people to be the "father" of the first PLC) said the following about the PLC: "It was even then a computer, but was not called that since it might scare off the user base, since they were not very 'computer literate' at that point."

I believe that for "PC based systems" to be better accepted, people will require a better understanding of what makes a system reliable. At this point though, the user base still isn't very computer literate.

In the past 10 years however commodity PC hardware and software capabilities have evolved and diversified. Making a PC based system that is equal in reliability to a PLC is becoming much easier and doesn't require exotic software designs. The decline of the PLC "rack" and the rise of Ethernet based networked I/O means that interfacing to industrial I/O has also become much easier.

I suspect that as a result of this, there will be a shift in the PLC market. For very small and simple systems, the PLC "brick" won't be going away. It's still very convenient and easy to install. For lareger or more complex systems though, the market will move to PCs.

These will not necessarily be immediately recognisable as PCs. They may be in nice plastic DIN rail mount boxes with all the usual brand labelling and logos from companies in the automation business. However, inside they will be PCs using commodity components. The rational for this will be the same as for the move to Ethernet based industrial networks. The cost and performance advantages will be just too big to continue to ignore.

As the hardware capabilities increase, so will the ability to take advantage of commodity software. As this happens, we will see things like web servers and databases becoming a standard part of any PLC. These will be used to store and analyse data relating to the performance of the machinery being controlled. The ladder logic part of a "PLC" will become a smaller part of the overall system. People in our business who fail to become more "computer literate" will find themselves being left behind.

The computer market evolved from mainframes to mini-computers, to PCs, and is now entering what I would call a "post-PC" era. This more or less describes the move from expensive proprietary hardware and software to smaller and less expensive proprietary hardware and software, to commodity hardware with proprietary software, and finally in the "post-PC" era to commodity hardware with commodity software. In the PLC market, we're still more or less stuck in the mini-computer era. I think though that the developmental forces are still the same in both markets. The evolution in the industrial market has just been slower.
 
C
Hi Michael I seriously considered porting a Linux PLC to the WinPLC. Seem quite doable since there is a Linux port for the H processor they use.  The Host folks, while friendly enough, had absolutely no interest in the project and the cost to do it on my own is prohibitive.

That is to say, there are cheaper alternatives for a sole proprietor or at least the cost is spread out to fit my toybox budget.  If someone has one they aren't using and would like to lend it or donate it, I am still interested because I still don't have a PLC hardware platform that runs Linux.  The Host folks might even come around if they saw something running.  The port would, of course, be  GPL.  I don't have the time anymore for a whole PLC system, but perhaps a stone for the soup?

Regards
cww
 
Top