Soft Plc

I'm doing a reseach project into PLC's. I've visited a huge Allen Bradley system and it was very insightful. I have 1000 words left to write and I'd like to concentrate on the the future of PLC control i.e SOFT PLC. Any ideas where I'll find 1000 good words?

The Future?

A soft PLC is not necessarily the future of PLCs. (But if I knew the future, I'd be out buying lottery tickets, or stocks.)

Keep in mind that a PLC is a Processor, a dedicated I/O scanner (at least for backplane I/O), plus a (firmware) operating system, which is dedicated to running the PLC program.

The hardware (processor) in a PLC is designed to operate under very harsh conditions. I've seen PLCs set up in unventillated boxes, outside in full sun (during summer, under snow in winter), which have been in continuous operation for more than a decade.

The hardware to run a soft PLC can be any old Intel-based PC. Now there are some very good PCs that have come on the market in the last few years which might be able to withstand the environment.

But every time I open my PC, and look at the FAN that MUST be on the Intel chip to keep it happy (let alone the one in the case on the power supply), I renew my opposition to soft PLCs (yeah, yeah, I'm just a stick-in-the-Mudd engineer.).

The main reasons to go to a soft PLC, according to the proponents, is cost. "You can get a PC for 1/10th of the price of a PLC". But by the time you harden the processor, and get a solid-state hard drive (i.e., no moving parts to fail), and get the I/O scanner (not cheap, and RS-232 usually doesn't cut it, if for no other reason, speed), then you are in the price range of a "real" PLC.

Then you add in all the other little annoyances ("The soft PLC runs in a kernal UNDER Windows. Windows has crashed (Blue Screen), but the soft PLC is still chugging along (=good) - how do I get online with the soft PLC, since that's done through Windows) without shutting down the line?") versus the benefits (more memory = bigger programs; faster clock speed = quicker execution), and the whole thing comes out to be a wash.

In my opinion.

(The above is over 1800 words, but is intentially not designed to write your essay for you, but instead give you topics/ideas to look at when researching soft PLCs.

Allen, as usual a great reply, but 2044 characters = 1800 words??

Not for my students!! But then that's in Canada where we now use metric. 10 bits= 1 byte etc.<grin>

Yeah, I slipped on the count. Phil has a "check message length" at the bottom of the Reply window. I clicked that, and saw 1833 and thought words instead of characters.

Now you know why my signature shows me wearing glasses.

Just figure out how to describe "How many times my PC has crashed this week" in 1000 words or more!

Industrial machinery can not/will not tolerate Bill Gates and Microsoft Windows.

By the way--How many Microsoft engineers does it take to change a light bulb???? None--Zero--they simply declare dark as the standard!!

One effect of soft PLCs.

A few years ago when soft PLC started to become a hot topic, the Modicon Quantums were just out.
We could buy a 186 based CPU with maths co-processor for just over $10,000 Australian. They were also quite unreliable and crashed slightly less times than out NT work stations.
Modicon, Seimens and Pheonix contact all brought out a soft PLC around that time. The main effect seemed to be PLC CPU prices dropping. The top line (586) Quantum is now slightly less than the old 186 based Quantum.

I think the main benifit to the the soft PLC was to introduce a competitor to the PLC market, with all the good things that competition provides.

For the future, I think that soft PLCs are viable alternatives to the high end PLCs for quite a few applications. I don't think that they will ever compete with the low end of the market, PLCs are just too good and too cheap.

Hope this helps to provoke a few thoughts,

Doug

There are definitely many good points made here. If you look at some of the claims that are made relative to PC-based control (I think the 'soft' in soft plc is intended to describe the ruggedness of the hardware platform) it seems that you get the ol' bait & switch. PC-based proponents talk about the low cost of hot processors due to the economy of scale. You can get a Dell 2GHz P4 computer for <$1000. But in the same breath they say they recommend a 500MHz industrially hardened P3 for >$3000 so the machine can survive the environment. And even if you do opt for the Dell you still need to load on a runtime engine for, lets say, $1000 a copy entry level. Then, as stated above, you need to buy an I/O interface card for about $1000 or so, unless you want to try your hand at Ethernet for industrial control. So cost isn't a real benefit.
Well, OK, what about leveraging the benefits of the Microsoft DCOM model to communicate machine information to everyone and their grandmother? Isn't that sweet? But you need to be careful because so many of those ultra-cool information exchange applications can cause issues with processor stability. So maybe you don't want to run those on the same box as the PC-based control engine or that 2,000 pound ingot you are moving around just might end up through a wall when the PC locks. But what about the real-time engines that run below Windows so a PC lock doesn't cause a control issue? That runtime engine costs $3000 a copy. OK, let's strike that benefit.
But what about the open nature of PC-based control? Who doesn't like choices? So I'm going to order a truck body from Dodge (I like the look), the powertrain and frame from GM (nice motors) and I'm going to go with Ford for the powertrain computer. And I'm going to toss it all together myself. So who do I call when the body doesn't bolt up to the frame or the powertrain computer runs into a transmission sensor incompatibility? It's not specifically anyone's problem but mine. So I get to play with it until it's fixed.
Now, I'm not overly negative toward PC-based control. I've used it in a couple of applications and it really is pretty cool from a speed and flexibility standpoint. But I also ran onto every one of the issues I stated above. Will PC-based control take over? Not completely. It's hard to beat a $100 micro for small machine control. And many of those can be networked now so you could develop a modular approach to control larger projects. But on the largest projects with motion requirements and high-end user interfaces I suspect PC-based control will start moving in. Then again, that ControlLogix thing is pretty fast and you can stick multiple processors in the same backplane.

So how did I do on word count?

Keith

A good solution has been offered by Omron. They have a real PLC on a PCI card. If the computer dies the PLC continues to run as it has a separate 24 VDC power supply connection to keep it running. It runs remote I/O on a twisted pair.
Beats the heck out of soft PLC. beerchug