Typical Performance between Ignition and AB SLC 5/05

I have inherited an Ignition Gateway and some clients. I recently got some Bad Tag quality issues. Approx 340 tags at scan class of 1000ms. I get a Device Loading of 650% with a typical mean response time of 500ms(seems super long)

I used a Trial Gateway on my laptop and added approx. 340 Tags and I can replicate the issue from my Production GW to this trial Gateway that I plug directly into machine.

Just curious how many tags people can normally read out of a SLC 5/05 at a scan class of 1000ms?
I’m working on moving these tags to a slower 3000ms scan class.
Also I am going to dig in on the Rockwell side to see if I can get better Ethernet Diagnostics on my SLC. I have a feeling I have either a physical Ethernet issue inside the machine or something with max Ethernet connections in SLC.

Others can comment about what speeds they see in the real world, but you probably have other systems polling this SLC as well, (prod gateway, your trial gateway, maybe RSLinx, etc…), and the more things polling it the worse the performance will be.

Check “Comms Servicing Selection” and “MSG Servicing Selection” for channel 1 (S:2/15 & S:33/7). If either is set, it will crush your throughput, as only one request will be processed per scan (Comms) or per certain control flow instructions (MSG).

6 Likes

As an anecdote, the rule of thumb was no more than three HMIs on a SLC. We have a line here that has 5 HMIs, four robots, three vision systems-- there may be a partridge in there somewhere. And they all want to connect to the SLC, or vice versa. The spec on older SLCs for simultaneous connections is 8. You can see where we may have a bit of a throughput issue…

Fortunately, this line will be going away in a few weeks, along with the headaches associated with it. While I always hate to see business leave (end of product life), I won’t be sorry to see the machinery go.

3 Likes

Thanks for the Tip on the S:2/15 bit! So I think I have one page on my HMI that bogs down my SLC Ethernet port and I can watch Device Loading go to 650%. Once toggled S:2/15 off my device loading unwound from 650% down to 30%. Also if I would have stayed off the one page on the HMI the device loading would have stayed reasonable.

I also turned on Ethernet Diagnostics but couldn’t see anything glaring. Turning off the S:2/15 definitely fixed the Device Loading/ Tag Quality issue. Now I just need to investigate the HMI and hope I did not introduce a Scan Time issue by allowing multiple comms to happen over 1 scan.

Ran into this thread trying to find a way to improve Ethernet coms to 5/05. After checking that bits S2/15 and S33/7 were cleared, I was still having issues. Adding a SVC instruction to the end of the main program greatly increased my throughput with little cost to scan time.

I have read in other forums of people placing them every 500 rungs, so far just the one at the end has been all I needed.

4 Likes

I can confirm this solution worked for me as well. Clearing S:2/15 and S:33/7 yielded minor increases in performance but adding a SVC decreased my overload from 170% to 60% for a 600ms tag group