Machinery Control. PLC over SoftPLC

[DRCDAZ]

The company I am employed by have been creating machines for web handling processes for many years, we have moved from "Relay" and "in house" control logic on to PLC's and intelegent drive solutions over the years. Our suppliers are now taking more and more about soft PLC's. At present we use PILZ stuff to handle all the safety side of things, and this would not change. My engineers and I are reticent to make the move to these soft PLC's as they seem a big leap, and we worry about reliabilty etc. We currently use a range of PLC's (Siemens, AB, Mits, SAIA, OMRON etc) but its seems that world wide we are being asked more and more for just Siemens or AB depending on the industry. Does anyone have any comments or experiance with these soft PLC solutions and how they rate performance wise against a standard say S7 PLC. Also any comments on the type of IO used with them?

 

[lexonrey]

A PC-based Control System can take the following configuration.

- A PC with remote / distributed I/O networks running a "Soft PLC" control application
- A PLC co-processor card installed in a PC backplane
- A PC co-processor card installed in a PLC backplane
- An open architecture PLC

I work on an Aluminium Smelter, with a Calciner plant too, all our process plants, we use an Allen Bradley RSView32 DCS software, running on normal PCs as workstations,using Control Net, an ethernet based system, connected to PLC 5's and remote I/O's that are addressable to the DCS level ( not dedicated to a particular PLC as it can be addressed on both ethernet and DH+ network levels). There are two servers, which is running all the time, but only one server on-line, the other server can take over immediately. Since we have a DCS server, we know we can connect it to the web, though we find there is no need.

So, far, after 3 years of operation, occasional server failures, but since another server is available, the effect on the plant is not evident. We just reload the server and away it goes again. Workstations sometimes locks up, like your normal PC locking up, but a reset always brings it up again. The Allen Bradley PLCs are robust and had not one failed so far, neither any I/O modules, remote or otherwise.

If the above can be called a "Soft PLC application " then I can tell you that the system is reliable so far, that is for our industry.

 

[ICman]

I would consider lexonray's system a Distributed Control System. The local control is thru PLC-5, with the supervisory portion thru RSView. The latter has redundant servers so that he won't lose monitoring or command capabilities.

"SoftPLC" to me indicates the normal functions of a PLC are performed in a PC. While a PLC contains a similar microprocessor platform, it is quite different than a PC. I offer some pros & cons (MY opinion):

* Even "industrialized" PCs cannot withstand the environment PLCs often encounter.

* PLCs are usually programmed in ladder logic, or another language that is understood by several. Using anything different might cause problems in maintenance and operation. They would have to learn a new language/method.

* PLCs are faster, if that's an important factor in your process.

* SoftPLC systems are not as well "integrated" with their I/O. Some pulse input cards have driven us fits, even when they are part of a PLC system. Using one in a generic manner is scary.

* If you have a problem down the road, you can usually get help from the PLC manufacturer. That would be more difficult with a "virtual" PLC.

OK, I admit I am biased toward traditional control. I did a study a few years ago, comparing traditional PLC control to PC-based systems. The PLC system won out by far. However, a lot of things have changed since then, so the comparison may not be the same.

 

[lexonrey]

Sorry, but now i do understand what the "Soft PLC" being referred to. We have not encounterd any pressure from our suppliers to switch to such system or configuration, so really have not heard of it.

If the DCS configuration is what the "Traditional" would be, a configuration that would cut out the PLC is not something we would embrace ourselves, immediately. But on the other hand, it is quite a good idea, not a new one at that, though. I/O cards have been available for the past fifteen years which could be inserted into the ISA slot on a PC. In fact, i have been encountering them at the same time when the Sequence Controller ( the first PLC i know, boolean programming too ) was in the Market. It is widely used for small and home automation control. If control system manufacturers are now offering the same thing, the level of confidence to such setup must have increased, the teething problems overcomed and reliability proven on pilot systems. Network the PCs and you have a DCS system, even a windows based ones. Historically, programming GUI available for PLCs had consistently followed the rule of user friendliness, even Allen Bradley PLCs have windows based programming interface, off a PC or laptops. Besides, we do protect the PLCs from the plant floor environment too, as we locate them in MCC rooms which are now with AC.

Come to think of it, we have used such "soft PLC" before on Kilns Firing control, actually to test pulsing gas solenoid valves to control the temperature. 2 ISA I/O card , one for Input ( 4-20 mA) and another for output, 24VDC,inserted into a 386Sx ( that long back ) IBM Pc ISA slot. We made a BASIC program and used that for 8 months on a set of Kilns Cells, and switched to PLCS and RSView32, having determined that it is more economical and reliable in terms of control strategy. Funny, we never thought of continuing such setup, as we applied the control strategy learned on the PLC. Of course, we were never interested if the pulses were consistent, as temperature control was the main consideration, and the survivability of the solenoid valves.

 

[DRCDAZ]

Thanks for your comments, it appears we have been correct in holding off, it was our belief that using a PLC for machine control is the better option over using a soft PLC, although the soft PLC can be often a more "cost effective" solution looking purely at the numbers :S. However speed was one thing we had not considered we had assumed similar loop times were possible. Just goes to show how some things can bee "Glossed" over by suppliers on occasions.

Thanks again

 

[QUIPPDUDE]

Product life cycle, reliability, ease of I/O connection, long-term support and deterministic control are key issues when considering PLS vs. SoftPLC.

LIFE CYCLE
With SoftPLC applications you are at the mercy of the constant upgrades/changes in PC's, system startup and maintenance are never straighforward. Troubleshooting a locked-up PC can be a nightmare.

RELIABILITY
How many times does a PC crash versus a PLC. Don't tell me your destop PC has never crashed, once is bad enough!!

I/O CONNECTIONS
Whether local [rack based] or remote, their are a multitude of options available with a PLC that include cross-platform gateways. SoftPLC's tend to be more limited in I/O types and cost per point is higher.

LONG-TERM SUPPORT
Ever lost or corrupted a windows driver. Try to find the original 18 months down the road?
Training - are you prepared to provide the support and instruction that will be required by the end user in all the nuances of a SoftPlc system in addition to the PC operating system.

DETERMINISTIC CONTROL
Potentially SoftPlc's are faster [4GHz processor!] but with the overhead of drivers, displays, USB ports etc. etc. they offer no speed advantage. The PLC offers a completely deterministic and predictable control system.

Well you see that I am biased towards the PLC, to date I am not convinced that SoftPlc offers ANY advantages.

 

[xnuke]

I also do not trust softPLCs yet, but here's a recent article about your question:

http://www.controlmagazine.com/Web_First/CT.nsf/ContentFrameset?OpenForm&ArticleID=DFUO-5LMTPB

xnuke

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[JonieR]

Just one more opinion: don't go there
plc's are comercial / industrial , pc's ALL are commodity junk!
all of the above negitives are true and more , been doing this for 30 yaers and tryed it all built written base software and all ..... no way not for control funtions.

 

[AGomaa]

I build system for industrial application, which requires 100% vailability @365 day/year.
I use PLC from Automationdirect they are way cheaper than the brands you mentioned. I could work out their PLC into a full redundant system with I/O over 500.
The only problem I found is the operator interfaces which in most cases does not last as the PLC do. Here I am mentioning years of operation.
My concern is the aftersales services and the time you may spend in less reliable system
good luck
Aly Gomaa

 

[DRCDAZ]

Again thanks to everyone for you comments. Just to let you know I am going to "trial" a Beckhoff soft PLC solution, they claim step times of nano seconds (or at least micro seconds)also they seem to bypass windows to remove some of the stability issues and claim the soft plc will still run even if windows has a blue screen of death. Although obviously there is the issue of how you interact with it if windows dies. However I can't get away from the potential cost savings so it is worth trying one to see what happens.

I will let you all know the results in the near future.