PLCs are going software? Soft-PLC vs Hard PLC

drbitboy said:

• Git
• OOP
• Web-based options

Click to expand...

Siemens can already do a lot of that "soft PLC" stuff with their hard PLCs. V16 introduced a "version control interface" that allows syncing PLC code to git/subversion/etc as XML. The PLCs have for a while been able to have custom web apps run out of the PLC, which can interact with it. The 1500's latest firmware enabled a web API (REST?) to interact with tag data, in addition to the OPC UA client/server already available. Siemens also has the 1518MFP, which is a hard PLC with a C++ runtime in parallel to the PLC code.

I'm sure some of these capabilities aren't unique to Siemens, that's just where my experience lies.


A lot of these features that are mentioned aren't incompatible with Hard PLCs, they just haven't been implemented yet because there is so little demand for them across the board. It's all niche requirements right now.

I don't understand why people think that Soft PLCs will be running on commercial PCs and are worried about the OS crashing randomly.
How many times have you heard or seen Twincat runtime or RT Windows crashing on Beckhoff controllers?
Soft PLCs are designed to run on Windows or Linux OS with real time patches installed and on rugged industrial controllers like Beckhoff, Wago PFC100/200 or Siemens 1500 CPUs that can be installed in a control panel.

Currently we are using a low cost Linux based Exor HMI which is able to run the HMI runtime, Codesys Soft PLC runtime, a VNC server, multiple industrial protocols, MQTT Client and UA Server simultaneously.
The PLC logic cycle time is extremely fast and much faster than when the same code is running on a Schneider Hard PLC. The jitter is also very low.
This just a low cost HMI with a single core ARM Cortex-A8 and 512GB RAM.
Never seen the OS crashed either.

mk42 said:

Siemens can already do a lot of that "soft PLC" stuff with their hard PLCs. V16 introduced a "version control interface" that allows syncing PLC code to git/subversion/etc as XML. ...

Click to expand...

Is there a usable equivalent of [git diff], to show incremental code changes? or is it using only the repository functions of git?

Bit_Bucket_07 said:

Seems like it was back in the late '80s that I first started reading about how the SoftPLC would make the PLC obsolete. I suppose it could happen any day now.

Click to expand...

Mid-80's, actually.

I always ask: beside <perceived> cost savings, what measurable thing is it about a PC-based solution that evaluates better than a PLC...I'll accept any list of items numbered starting with 1.

What about the electronics - PLC's use higher rated temp chips have redundant and surge protection some even are mil grade certainly industrial grade - soft PLC's powerful as they are have there place (Example Injection moulding machines using a robust bespoke OS for the control and windows OS as the HMI) but hard PLC's also have there place especially in high speed control dependant situations.
Both work great side by side one will not overtake the other similar discussions about PLC, PAC and DCS the industry will always use the best component for the job at the best price.

drbitboy said:

Is there a usable equivalent of [git diff], to show incremental code changes? or is it using only the repository functions of git?

Click to expand...

Honestly, never had the opportunity to actually use it.



Poking around the support site indicates that the Git plugin only has repository type commands, but there is a SIMATIC Automation Compare Tool that specifically indicates it can compare the code in your project to git (and translates the XML back into something human readable).

This topic comes up every year or two. Nothing will change that rapidly.
BTW, PLCs are firmware. Whether you want to admit it or not your glorious 2000 rung program is only 5% of the total project. The rest is in the firmware.


Delta Computer Systems has made motion modules for different PLC platform. PLC companies have been to the Siemens lab in Johnson City, the PAL ( Prouct Assurance Lab ) for Modicon and the Rockwell testing lab in Mayfield Heights, OH. I know they all test their products much more thoroughly than a PC manufacturer would.


There are three classes of components. Commercial, Industrial and Military. Most PLC stuff is using commercial grade chips to reduce costs. They just aren't pushing them as hard as a PC manufacturer would.
Heat dissipation is a big problem.
Just because a control system is PC based doesn't mean it is not industrial strength. We have used harden PCs from Advantech. We run a real time linux on them. The combination has been reliable. In the same system we also use Rockwell PLCs for handling the I/O and timing. All of this supervises 64-96 servo motors.


Everything has advantages and disadvantages. What really seems to make a difference is the staying power of the supplier. I have worked with Steeplechase and Think 'n' Do. Where are they now?


The PLC manufacturers that make industrial products have nothing to fear from "soft" PLCs. However, there are always niches.

JesperMP said:

I do think that soft-PLCs will become gradually more prevalent.

The article lost me immediately with its silly comparison of Kodak and Film vs. Digital to Siemens/RA/ABB/Schneider and Soft vs Hard PLC.
Digital imaging have big genuine advantages over film that made it disruptive. Digital imaging enables something that was impossible with film, i.e. instantly transmitting images to be used in media with no delay. With film, you had to wait for the film to be processed.

What are the similar advantages that soft-plcs brings to the market that cannot be done with traditional PLCs ?

This is an example of how trying to force an argument by using a poor analogy sets you up to be counterargued easily.

Like I said, I do think that soft-PLCs will become more prevalent. But this article is awful.

Click to expand...

The article makes very valid points, whether the Kodak analogy was good or not. They even present data, with at least one soft PLC manufacturer grabbing market share faster and faster, while the other hard PLC manufacturer is flattening out over the past 5+ years. Don't ignore the data. RA would be foolish to ignore the data as well. However, I believe that is mostly what they are doing right now as evidenced by them not being very innovative.

It's just as the article states, most companies won't jump ship from whatever PLC platform they are on simply because it's too expensive in downtime and training, and X-brand PLC is what their people know. If it wasn't for those things, hard PLC's would have been gone 10 years ago. I don't think its a matter of "if" the hard PLC is displaced. It's simply a matter of "when". Just like it's just a matter time before the EV displaces the ICE in the automotive landscape. With that said, I don't think the soft PLC will completely displace the hard PLC within the next 20 years. 30-50 years out? Absolutely.

Have seen all this before. Currently in Oz everything has to be dual redundant including racks. I/O has to be a ring. How is a soft PLC going to handle that? Schneider have done a very good sales job on frightening engineers to sell there very expensive hardware and software.

BobB said:

How is a soft PLC going to handle that?

Click to expand...

For starters it's easier to add network ports on a PC/Server than PLC's. And the redundancy is software... a soft PLC can run on a redundant server (like the Stratus kit) where the software is none the wiser about the hardware underneath.

This isn't a new contention. It goes back at least to the '90s when industrial PCs first started to hit the market. The tough, reliable, adaptable PLC will continue to hold a big share of the industrial automation market for a long time, in my opinion.