An Allan Bradley 5/04 PLC program will scan 200 lines of code in X millisecond.

What if that program has analog input modules, counters, timers, and is interfaced with a Dataliner? Want the scan time increase? How can one determine how long the scan time will be?

 

Download the SLC instruction set reference from http://www.ab.com. There is a table in the back that will help you estimate you scan times.

 

I prefer the experimental route. After the program is written and working, add one rung of ladder that looks like this:

--|/|-----()

Physically wire the output you're firing to the input you're examing if open. This will toggle the output every other scan. Now hook an oscilloscope up to the output and watch the rate at which it changes states. There's your real-world scan time.

-James Ingraham
Sage Automation, Inc.

 

If you're using A-B kit, then the real-world scan times are shown in the system register, under 'scan times' tag, (for non-logix processors) or in the properties of each program for the logix series.

I guess, however, that you are more interested in predicting then scan time before the system is built; then the calculations mentioned in the first reply is your best bet.

Incidentaly, I did some testing on the Control-Logix processor, which did 2000 lines of combinational logic (10 elements per line) in 150 MICROseconds (not milliseconds!) and 4000 lines in just 170 uS! The moral? If you are worried about scan times, think about migrating to the Logix range.

regards,
Steve Jackson,
Routeco Plc; UK

 

If you really want to make your CLX fly, upgrade to firmware 11.11 and use ST. Ive used ST on PLC5/xxEs and for one short bit of code on a CLX. I thought it would be slower or at least the same, but an assoicate benchmarked it on a Logix5550 and he tells me it blows LL out of the water on speed. Go figure,..... Like LL, it is still interpreted. Disadvange: you can't see it operate, so maintenance techies around here hate it. Love that CLX though, everything new from now on is noting but CLX platforms. Last project we had two programmers doing simultaneous online programming and the CLX kept it straight.

 

Hello List,
Hello James,

It does not show the worst case of program response time...

1. the worst case of the "execution path" must be defined (it is pure analytics);
2. the response time must be measured (or calculated).

Theoretical possibility exists, but, alas, I never see the examples of the implementation...

BTW, for the OSes kinda QNX or MS-clones implementation of the features is too difficult (imposible) because of unavailable sources/or formulas.

--
Best regards.
Vladimir E. Zyubin mailto:[email protected]

 

If you are really concerned about scan time then you should be looking at the Modicon Momentum Ethernet Processor with I/O on Modbus Plus.

I did a natural gas separation plant the first of the year with this processor. We controlled 29 loops, pressure, flow, level and temperature, with the loop solve time as fast as the scan. Along with all of the furnace control and pump control. The SCADA was Intellution FIX32 over Ethernet. The scan time was 43 milliseconds.

Using state logic, the scantime is really null because it only looks at the next state, not the entire thing.

It would be interesting to see the performance of Logix running a process with floating point loops. That would make it breath hard.