Life span of a PLC?

[Andybr]

It is easy to look upon systems like the PLC2 or S5 as being a bit antiquated but we have several Control Logix systems which are over 17 years old and I do not anticipate replacing them any time soon. If an old processor fails I can usually just replace it with a newer version. the only real issue is the lack of a serial port on the new models. Our oldest PLC's are currently Mitsubishi F1's which must be around 27 years old. If one of these fails the small IO count and limited functionality makes replacing them with a small modern PLC fairly simple. The programme can usually be rewritten before the replacement hardware is delivered. I am not saying that you should not plan for replacement but losing an old PLC is not that difficult to cope with.

 

[ASF]

In some ways I'm inclined to agree with Andy - with the CLX platform being 17 years old and still the flagship model, I feel like we're starting to get to a point where we might have to replace/upgrade CPU's and communication models from time to time, but that the system itself has a lot of life in it yet.

Another part of me goes "but if you have a 100-year-old axe, and you've replaced the handle twice and the head three times, is it still the same axe?"

And then another part of me goes "you're being hopelessly naive, technology is moving faster than ever, everything you're designing today will be obsolete in 10 years"

So I'm really not sure. If I had to throw out an answer, I think 20 years is a good bet too

 

[harryting]

Depends on how close the PLC is to the event horizon of a black hole.

 

[Pete.S.]

I am not saying that you should not plan for replacement but losing an old PLC is not that difficult to cope with.

True, if it's small system but there are a lot of big PLC installations running on hardware from the 80s.

I replaced a Modicon PLC system installed in the 80s. This was mid 90's and they had two man-years of development put into that Modicon system. It was complex and had lots of I/O and I think they'd maxed it out in every possible way.

I've worked on PLCs from the 80's that have more then a thousand digital and analog inputs and outputs spread over multiple remote I/O racks.

It's not an afternoons work to replace something like that.

 

[CapinWinky]

I would expect a system with a panel that is well protected to last for 20 years before components started failing for reasons other than manufacturing defects; however, as the article posted by JordanCClark points out, the lifespan of the components should not be driving your upgrade. I mean, the last thing you want is your upgrade to be driven by a failure and line stoppage. You want to upgrade on your own terms and with minimal/planned downtime.

I regularly see 30 year old systems in the field and even wire their discrete outputs to a new distributed IO island so a newer central system can remotely control them over Ethernet. In my particular case, there are newer versions of the machine that are roughly the same mechanically, so if an old system failed, it could be refurbished in the field with a new control system pretty quickly. That doesn't stop the end users from buying IO cards or sending them out for repair for insane prices. They can't repair/replace the PLC itself because they don't have the code on hand anymore and it would probably take as long to get a laptop running the old programming environment, dig up the program, and get the settings right as it would to refurbish.

The more successful end users I see are not penny wise and pound foolish enough to keep an old system running the same way for 20 years. They upgrade the system capabilities to improve performance, usability, and efficiency every 8-10 years. This is also frequently enough to have guys that know the old stuff and the new stuff in detail to make the conversion easy. Beyond 10 years, too much has changed and too many people have quit or been promoted to find anyone that knows both the new and old.

 

[Stephen Luft]

Pete,

As has been conveyed by previous responses, there are restrictions on the components used within the PLC. For example, mechanical relays used for outputs will have a rating of anywhere from 3 million to 10 million mechanical cycles (no load). When a load (amperage) is factored in to the operation of the relay, the cycle rating usually drops to approximately 100K and will vary depending on the amount of the load being switched each time. The greater the load, the shorter the anticipated life.

Other factors, aside from operational frequency that will impact the longevity of a PLC are ambient temperature, vibration, EMI, and RFI.

Solid state controllers will tend to last longer than relay output controller in that there are no moving parts that will wear out over time.

We are still replacing controllers from the early 80s which was the infancy of Entertron. This indicates that controllers over 30 years are still functional and in use. I can't say if this is the same with every manufacturer, but this has been our experience. Furthermore, I can't say if you will see the same longevity with newer products today than those that were produced over 30 years ago.

 

[mk42]

BTW, how long does PLC manufacturers promise to keep spare parts around?

Not sure about other vendors, but Siemens generally promises 10 years from the date they announce the product cancellation, at least for core automation products. They products are usually actively sold for a couple more years after the cancellation is announced, before they become spare parts only. Things like PCs have a shorter lifecycle.

From what I've seen, they mostly stick well to it. Except the recent HMI migration, where they can only sell spare parts with a new part # that isn't a drop in replacement.

 

[splicer480]

I think this is a critical line from that article too.

"Ask yourself: Do I really want to rely on spare parts from eBay to keep my plant in production?"

I wish I could get this point through to upper management.

I would like to upgrade several machines in our facility. Number 1 to run smoother and number 2 to be able to get quality new working parts instead of, "Well, lets try this thing."

 

[gbradley]

Depends on how close the PLC is to the event horizon of a black hole.

And we've got a lot of runway left before that happens.
timeline-of-the-far-future

 

[rootboy]

Depends on how close the PLC is to the event horizon of a black hole.

Inside, or outside of the TARDIS?

 

[kwade]

I wish I could get this point through to upper management.

I would like to upgrade several machines in our facility. Number 1 to run smoother and number 2 to be able to get quality new working parts instead of, "Well, lets try this thing."

I tell them, "It will schedule itself!"

I have about 40 SLC-5/03s and 6-8 SLC-5/04s and as about many PV550s from, what, mid/late-nineties still running without too much headache.
My pucker-factor increases on a daily basis. The lines have had a great run and are too old to spend much money to fix/repair.
I can't afford a $200+ backlight for 30+ PV550s. As an interim fix, I designed an LED controller around a Bridgelux BXRA-56C0700-A.
It fits the backlight lamp holder nicely. It is ON/OFF/Brightness is controlled with an Atmel ATTiny84/IRF510.
The original lamp power wire acts as the enable to the lamp controller.
I have requested that we re-design the machine Monitor/Control circuitry around a new PLC.

 

[FactoryTalktotheHand]

Infinite.
Relays, Capacitors and LED's have a limited lifespan. Power supplies, HMI's, and VFD's will be the issue. As long as replacements are available, you can run forever.

Well, technically yes, but you have external factors at work here as well.

For example, the other day I had to go online with a PLC-5 and the only port available was DH+. Now, software-wise, RSLogix5 will install on a Windows 7 machine so that's not an issue, but the problem is hardware. Modern laptops only come with half-depth PCMCIA slots, so the 1784-PCMK card can't plug into them. The only way to get online DH+ with a modern laptop is A) buy the $3,000+ USB adapter from Rockwell, or B) lug a Control Logix rack with a DHRIO and ENBT card installed in it to the job site. The only other option is keep an old laptop around with a full-depth PCMCIA slot and the hardware/software.

The one laptop we have in the company that can get on a DH+ network belongs to a technician who just retired. I had to drive to his house to borrow it. Now, I did eventually get online and troubleshoot the machine. Heck, if I really wanted to I could hook up to a PLC-2 or SLC-150 if I needed to. But guess how much time I spent hunting down all the equipment I needed to get online with that PLC-5? About half a day. And I'm billing that. And this situation is only going to get worse. Once that laptop goes kaput for good, we will probably have to start telling customers with 30-40-year-old PLCs that we just plain can't help them. As time goes on, we won't have the knowledge nor the tools to do so. Even in cases where we do, the customer is going to pay more to service that machine, because we have to spend more time digging up old hardware and re-familiarizing ourselves with the software.

Hardware life is only one part of the equation when it comes to obsolescence. Serviceability is the other, and as laptop technology advances and changes, it's going to be a struggle to maintain support for legacy systems. You can theoretically keep a car running forever if you simply keep replacing parts, but there comes a point where it's no longer financially practical to keep it running as a daily driver. At that point, you either junk it or continue putting money into it as purely a labor of love. And I don't see a lot of industrial machinery kept around to serve as museum pieces.

 

[geniusintraining]

And we've got a lot of runway left before that happens.
timeline-of-the-far-future

Thank GOD I went and saw Niagara Falls last month.... its going to be gone in 50k years