"Recipe" handling system woes with SLC and FactoryTalk View

cardosocea

Member
Join Date
Nov 2016
Location
Fields of corn
Posts
2,622
Hello Guys,

I've inherited a system that manages the bulk material loading onto 4 reactors. The system is built on an SLC PLC and the majority of the valves are on a DeviceNet network where we can issue commands and receive their position feedback.

Apparently this has been unreliable from the start and although we are looking at the entire replacement of it, with the Covid-19 now appearing we're trying one last push to get it working properly and potentially extend its life a bit longer.

The problem we have is that the additions every now and then either go completely wrong (wrong material added) or to the wrong location or the quantities are off (like double charging).

The logic in the PLC essentially waits to see all the quantities to change from -32767 to something different and if all of them are different and a start bit is made, it copies the addition data onto a buffer and then another part of the program deals with the requests in the buffer and routes the addition through.

I've followed the logic and can't quite see what could cause this sort of problem there as it's not consistent, so gut feeling told me perhaps the problem is in the SCADA (FactoryTalk View SE).

In there it essentially takes the values for each addition and writes to the tags in the PLC. This however is done with the Set command and, according to the manual, to speed it up they used the & to execute these in parallel fashion. This is one thing I wasn't sure would be the cause of the problems, but since the logic waits to see all the variables changing (by checking for bit 15 to go to zero), I assumed it would be ok.

I know the description is quite vague, but would you guys have seen any similar system before or potentially problems that would explain these woes? Or places to look for what could cause this too.

Likewise, the pop up from where the additions can be controlled is available in several nodes. Is there a known issue when using reusable pop ups across different clients for the same application? Meaning that we have 4 reactors to load and the system looks at all 4 so potentially two additions may be asked to be performed at the same time. Would there be a way to limit this?
 
without more details it's hard to say
But it sounds like the recipe is changing while the processer is weighting
maybe you are righting to 2 different recipe's at the same time
try not editing anything while any of the 4 are running
you could also try buffering the date from the Factory talk View the with another button load it into the batching system tags
If you are planning on upgrading I would recommend that you go with a Compact Logix processer
also Devicenet is being phased out so consider upgrading the network devices as well
 
Yeah, I'm not a fan of HMI logic myself.
As for the recipe changing, I don't think that is the case in all the instances as a lot of times, the wrong material gets added, so the error happens before the weighing begins.
How would I avoid the different FTView Clients fighting to send their recipes? If one terminal has the recipe pop-up open, the others wouldn't be allowed to do it?

The devicenet is going to be an absolute monster to replace since all the valve positioners and actuators are on deviceNet. Rockwell not having a proper fieldbus at the moment and worst of all not a lot of companies doing networked valves other than on Profibus, means that I have to go down the IO route. Adding IO for all of them valves is not going to be a cheap. I did wonder whether it would be some quirk in the DeviceNet that caused this problem, but it strikes me as odd that a quirk would work so nicely with so many valves, so most likely not.

Anyone here has used networked valves? The current system I have is the Westlock Intelis.
 
I have used Burkert on Profibus, we had something like 280 + valves on 24 way islands and Siemens 400 PLC, these were incredibly reliable in the 22 years of service we only had one valve fail.
 
I have used Burkert on Profibus, we had something like 280 + valves on 24 way islands and Siemens 400 PLC, these were incredibly reliable in the 22 years of service we only had one valve fail.


Yeah, I think as a way to save a bit of upfront cost we may keep the deviceNet system as it's still supported.


Go over to an S7-1200 or 1500 and use a Profinet to DeviceNet gateway like the Anybus

https://www.anybus.com/products/gat...ateway-devicenet-scanner---profinet-io-device

The SLC is pretty bullet proof as is DeviceNet so its a programming issue and from what you say probably in the Scada, shouldn't be that hard to fathom out


Sadly, the company decided to standardise on Rockwell, so no Siemens for me. Though as I said above, I think for the time being we'll keep the existing valves as they are since there's still support for them (though I'll need to make friends at Westlock).



I'm sure it's a programming issue but it hasn't happened since I joined. There were about 6 events in 4 months (and the system is used daily).
 
Go over to an S7-1200 or 1500 and use a Profinet to DeviceNet gateway like the Anybus

https://www.anybus.com/products/gat...ateway-devicenet-scanner---profinet-io-device

The SLC is pretty bullet proof as is DeviceNet so its a programming issue and from what you say probably in the Scada, shouldn't be that hard to fathom out

DeviceNet is what?

I think DNET and CNET are both garbage, yeah great when they work, but zero tools when they suddenly stop working. No thanks.
 
The alternatives being?

Ethernet, its used by millions of IT groups and they have all kinds of tools for it. Probably the biggest issue I have with Ethernet is potential duplicate IP addresses and multi-casting. If you set it up correctly (switches/network) it works well. When it stops working I feel figuring out the bug is easier.

For example, I had a core switch that was going to a dual vlan NTRON that was going to another dual vlan NTRON that eventually made a loop to cisco then core switch.

Someone moved (relanded the port) some ethernet "fixing" something and this was making the network have random failures. When I tracked it down I made the NTRON single vlan and removed the loop. I was able to do this with limited IT support or tools. I fixed this in a day start to finish.

If it was CNET or DNET and it hypothetically randomly drops what do you do (I'm actually asking because the more tools I can get the better.)? The networks I was working with span a lot of systems and are in conduit with drawings that may or may not exist.
 
Last edited:
The question was in relation to my existing application... yes. Ethernet is the way things are headed, but very few other than high value devices actually interface with Ethernet. You get drives, IO blocks and flowmeters, but if you want Ethernet controlled valves you're stuck (or certainly I am since my searches came back empty).

I must admit I'm not the world's expert on DNET or CNET (I've mostly used Profibus before) but at this moment replacing all of these valve actuator/positioners with hardwired ones plus IO blocks will be an absolute fortune both in hardware and in labour to run and isntall.

Do you know of any manufacturer that does such an actuator system for ball valves?
 
My first impression is that you can make the system significantly better with keyboard magic. I have had a lot of experiences tiptoeing through existing code making small repairs until eventually the whole thing was pretty much my own.

I have at least 5 times after the fact said to myself "I should have just wrote a whole new program from scratch. It would have been faster and easier".

That last statement does not apply to machines and their associated systems that I don't thoroughly and completely understand. Sometimes you have to do piecemeal mods to avoid causing downtime on critical equipment even though it is more work and the end result is not as clean.

The only low cost actuators I have seen that support a fieldbus were Modbus RTU capable. I think that might be a step backward from DeviceNet, so I would probably not go that route, and if I did, there would be a 3rd party comms handler in the middle of it.
 

Similar Topics

Just a little bored, so I thought I would start a thread in one way of using recipes in batching processes, I'm aware there are probably many ways...
Replies
0
Views
560
I'm new with the SCADA System Citect (V6.10) and I have the following questions. I need to create a recipe in dbf format with 100 variables. I...
Replies
0
Views
2,325
Hello. I was wondering if anyone had had a need to 'roll their own' recipes? My customer has a SLC 5/05 connected to a PV+1000 all on ethernet and...
Replies
0
Views
2,011
I'm using legacy recipe to download a recipe to the PLC. A little background, the recipes are large. They are broken down into 6 tables, each 16...
Replies
2
Views
80
Hi guys, got one problem PLC model fp-xh c60t and HMI model TG765S-MT I've encountered an issue which is the recipe file went missing if i swtich...
Replies
0
Views
76
Back
Top Bottom