Thinking through control hardware for upcoming project.

I'm continuing to think about choices in an upcoming plastic extruder project, and I'd like to talk it through with you folks. So far, the main control components that I'm considering for the machine are:

PLC: CompactLogix (not sure which model yet) This replaces the original SLC 5/05.

Temperature Controllers: Eurotherm Mini 8, probably 4 of them (approx 32 zones), Modbus TCP

Pressure controllers: Eurotherm 3504, 2 of them, Modbus TCP

VFD: Yaskawa GA800, 2 of them, Ethernet I/P (we prefer using Yaskawa VFDs)

HMI: Ignition Edge, possibly using OnLogic fanless NUC (not sure of sizing).
Touchscreen: Possibly Automation Direct, Atlas Industrial Touchscreen. A little worried about supply.

Any thoughts on these choices? At first I was looking for other temperature control options, but I'm leaning toward Eurotherm products to simplify their back-up parts, since they are the same as the last machine I did for them. I wish I knew more about Eurotherm's (Now Schneider Electric) future support for these items. As a tech, I constantly have battles with obsolesce. The cost per zone was a reason why I didn't go with analog input cards on the CompactLogix.

I'm also thinking about having the Eurotherm devices communicate with the PLC instead of the HMI (last time the HMI went directly to all temp and pressure controllers). I haven't tried it yet in my shop, but I suspect it isn't too difficult. Ignition Edge comes standard with 2 controller connections, and charges for additional connections. Going through the PLC for PV, setpoint, and output would seem to make sense to avoid the additional connections. I just can't think of the downside of doing it that way would be. Any thoughts?

(I'll probably copy and paste this post to the Inductive Automation board, as well, just to get more familiar with that resource.)

I work on a lot of plastic extruders. Overall, I like the way you are leaning. When I started upgrading our extruders, we used temperature cards, discrete input and output cards and current monitors to close our temperature loops. We used the PIDE blocks in the PLC (ControlLogix) to control temps. Our current setup uses Gefran GFX4 controllers because they do all of the things that we needed without all of the separate I/O cards and the PIDE programming. The GFX4's use E/IP for communications. All that being said, you should be able to accomplish the same things with the Eurotherm controllers (not sure about the current monitoring, though). The VFDs are solid. I can't argue against the hardware you selected. You would know what hardware serves your location the best.
Regarding Ignition, I have not used it yet. I need to make myself try it because of all of the great things I am hearing. The main thing I wanted to comment on is that I firmly believe that the HMI should be an interface only, not a communication bridge. By that, I mean that all control components should go to the PLC and the system should not depend on the HMI to keep running. Of course, if your HMI PC fails, it will not be easy to run the machine, but it will still run. In my experience, I have had more PCs fail than PLC components. My thought is that if a HMI fails, I can keep production running by connecting a programming PC to the PLC while we work on configuring and installing a new PC. If you connect your Eurotherm devices to the HMI PC, then they will quit if the PC fails.
By having everything connect through the PLC, you will also have an easier time getting access to all of the tags in each device for the PLC to read and write.
Those are my opinions.

+1 to have the devices connect to the PLC, instead of using the HMI as a bridge. See if they have an EIP flavor though, for easier integration. If not, then RTA makes a nice unit for MODBUS/TCP stuff.

https://www.rtautomation.com/product/460etcms/

For the NUC the CL210G-11 is plenty for Edge.

A quick google search... the Mini8 has an EIP option card available (yay!). The 3504 doesn't mention EIP... "Modbus Master, DeviceNet, Ethernet, Modbus, Profibus". I'll keep thinking about possibilities with that one. drforsythe got me thinking about Gefran on my last post about this. I believe my customer uses their pressure sensors, so it's probable they have a controller solution.

Thanks for the CL210G-11, the_msp. They are out of stock, but I'd like to at least have one in my shop to play with. That's not a bad price to pay for that. Too bad they are out of stock, though. I'd get one right now.

Well, my experience with plastic extruders is limited. But examining your list of components, from what I can tell, you can accomplish it with what you have there. But it appears to me you could easily do it more simply and for less cost.

Your choice of CompactLogix - Obviously an upgrade over a SLC 5/05. But why not look outside the conventional "black box" PLC and take a look at cutting edge automation technology with PC control?? Yes, you can code it in ladder. I would look at Beckhoff's CX2040 series. I would bet the CX2040 would easily surpass performance of whatever CLogix processor you're considering, and for less $$.

Eurotherm controllers - What advantages do these bring? What can these do that the PLC cannot? It seems like a lot of extra unnecessary cost. If I were using a CX controller as mentioned above, I wouldn't even bother with these. No need. Whatever interface and manual controls these have, you can integrate into your HMI. The PID control can obviously be done in the PLC.

Yaskawa VFDs. No questions or concerns here. Yaskawa a good name in VFD's. With that said, if the VFD doesn't have the option for EtherCAT, then it's a no-go in my book. I believe Yaskawa offers EtherCAT for their VFDs though. We standardized to Rexroth, another really good name in VFDs.

HMI, Ignition Edge: Better options out there, and for a whole lot less $$$. For instance, just one option of several is AdvancedHMI. It's free, plus you'd have the power of Visual Studio to do all sorts of cool things with the HMI. I prefer writing my own HMI, but we have also used AdvancedHMI on a few machines in our shop for it's "canned" HMI ease of use. It offers a very nice compromise and mix - canned HMI software for ease of use, but power of Visual Studio to do custom things you might want/need. If you go with a CX as mentioned above, you can have your HMI software installed on the CX right along with your PLC program, 3rd party software, etc. The advantage of this is that there is no communication to/from a PLC over cable, which gives you a very fast and responsive HMI program. For instance, the analogue values on our HMIs (with software installed on the Beckhoff CX machine) updates every 10ms, which is actually faster than the refresh rate of the monitor. So analogue gauges on the screen are smooth moving and very little noticeable lag, if any.

I don't deliberately mean to go against the grain on some of your choices. I'm just looking at it if I were given this project to build, these are questions I would ask, and options I would look into.

busarider29,

he already knows the ab software and there is a conversion tool from the slc to compact logix program. it's not a 100% conversion on some commands, but close.
if he does not know Beckhoff, he has to buy the software, developers and runtime, PC, all the components, and learn the software. then there is the question of does maintenance know the software, training them, learning curve.
regards,
james

A couple of quick thoughts about the last post by busarider29...

Beckhoff: Someday I'll do something with a Beckhoff system. Back in school, we covered them right before graduation (May 2010). They are local to Minnesota, and one of my classmates went to work for them after graduation... I believe she went to the tech help department. That someday won't be for this project, for reasons I won't get into here. (I also own Studio 5000, and have a tech connect automation contract.)

Eurotherm: The cost per zone is below the PLC cost per zone.

Advanced HMI: I've played around with it. I even used it for a Modbus communication lab in my shop. Quite frankly, Ignition is going to be part of the future, there's no doubt about it. I very much want to get some experience with Ignition, even if it's just the Edge version. This is the perfect project for jumping in... the customer completely trusts my judgement for picking the components that are right for both him and me. Ignition is the right choice for his needs, and mine.

CompactLogix sounds like a logical choice. I would use a 5380 series like the 5069-L306ER because almost everything we do uses two IPs, one local I/O network and one for the plant network. If you need more than 16 ethernet nodes or have a large program, step up to L310 or higher.

I am unfamiliar with the Eurotherm controllers. I briefly researched them. It looks like Mini8 does Ethernet/IP so they could be put in the I/O tree without Modbus. It's too bad Ethernet/IP doesn't seem available on the pressure controllers. It would be nice to just put them in the I/O tree and skip Modbus. I would be tempted to look for something else that does Ethernet/IP or consider doing those in the PLC. I wonder if you could find something suitable with Ethernet/IP from RLC or Watlow.

If you go with the Eurotherm pressure controllers and use Ethernet/IP on the Mini8's, I would probably put a 5069-SERIAL in the PLC rack and do Modbus RTU to the controllers. I would do one on each port since it's two ports for two controllers to keep the wiring clean. I found the 5069-SERIAL easy to set up, with everything under the module definition. I think the tag structure created is nicer than the array of data from using a ProSoft module. I haven't used RTA yet but I expect similar. It keeps everything Rockwell without needing the third party device. Using RTU, you might be able to cut controller costs if the 3200 series does everything else you need. Again, I don't know the controllers.

I am a big fan of the OnLogic fanless units. We use them extensively with Hope Industrial touch monitors which I also endorse.

Unfortunately, many things are in short supply so whatever is ideal might not be available. Good luck finding everything.

Thanks for that post, 5618. You filled a couple of gaps for me.

Yaskawa VFDs are still my favorite, but I am not hugely impressed with the GA800 series. We've had two failures out of the first dozen or so we have installed. They just aren't built as well as the older models. I appreciate the fact that they pretty much kept exactly the same parameter naming and numbering as the older models and that the Modbus RTU works exactly the same since that makes it easy for those who have canned code or familiarity to migrate. Their performance aside from the two duds has been great though.

OkiePC said:

Yaskawa VFDs are still my favorite, but I am not hugely impressed with the GA800 series. We've had two failures out of the first dozen or so we have installed.

Click to expand...

Yikes! We are an authorized service center for Yaskawa VFDs, and as a company that specializes in VFDs, Yaskawa's quality and support for legacy drives is a big selling point. I've done a couple of GA800s, and recently a couple of GA500s. And next week, I'll be retro-fitting a couple of 100 HP GA500s at a wastewater plant. Hopefully, we'll avoid the lemons that you've experienced.

I cant help but think to myself that this is yesterdays way of doing things.
The only reason why I would ever consider an external controller would be if it could do some magic tricks that a PLC cannot - fast motion controllers possibly. Temperature processes are slow. PLCs can handle them effortlessly.

It is not a matter of whether or not the PLC can handle it, it all comes down to money for our company. I can implement 16 channels of electric heating with cold water cooling (32 temperature zones) significantly less expensive with separate controllers on E/IP than I can using PLC cards. When I used to use a PLC for temperature control, I had the following for each zone:
- Heat enable relay [Digital output wired to control relay capable of handling the current of the electric heater] (it can be shut off in the event of runaway or temperature outside limits)
- Solid state relay [Digital output wired to SSR that would provide power to the heat zone]
- Thermocouple input [Thermocouple input on PLC]
- Current monitoring relay [Digital input to PLC] (tells control system when current is being drawn on a heat circuit. This would identify 1) if there was a problem with the heat enable relay, a fuse or the heater element or 2) if current was detected when not calling for heat, then a bad SSR or PLC output.
- Cooling solenoid [Digital Output from PLC]

Now that I use the Gefran GFX4 controllers, I just need the heat enable contactors, a GFX4 master and three slaves for 16 channels of heat with associated cooling zones. I send my setpoints (and all other settings) as well as read status (temperature, alarm, control state) via E/IP. Current draw on each heat circuit is measured within the controller and reported via status register. An added bonus of doing it this way is a significant panel space savings.

I realize not everyone uses enable contactors and current monitoring. If you do not consider those in your designs, then the PLC is probably the way to go. I like the added diagnostics of monitoring the current and having the enable contactors to shut off in the case of runaway or component failure.

Learning curve is a legitimate concern, but nowadays you would have to teach college students LD. Most of them are pretty comfortable with C style programming. By the way, you don't pay for Beckhoff software. You can download and run Beckhoff TwinCAT for free. Just keep renew the 7-day trail license.