Handling Stops & Resets in Sequential Programming

[DamianInRochester]

I'm working on a PC based control system now (Beckhoff/TwinCAT). I am programming it primarily in ladder.

 

[kolyur]

Back to the sequencing topic... I've also graduated to the numerical sequencing method as others have mentioned. The only advantages I can see to bit-based sequencing are that it's easier for B*bba to troubleshoot (maybe) and that you can have parallel threads within one sequence. That is, the sequence can diverge by activating two step bits simultaneously, then later converge when both separate mini-sequences are complete. This has some benefits but in retrospect I've found it more difficult to troubleshoot.

With numerical sequencing I've gotten in the habit of using steps 10-90 as startup steps, or things that need to happen when the machine is powered on or reset. Step 100 is the "ready" state, indicating that startup is complete and the sequence/station is idle. When the sequence is triggered to start, I move 200 into the register and the steps progress from there in increments of 10 (allowing room for expansion), as high as necessary, before returning to the idle [100] state. The nice thing about this setup is that you can quickly look at the integer data file (I usually create one just for sequences) and see the condition of the entire system. If the machine is ready to run, they should all be 100's. A stuck condition can easily be located by looking for values >200.

I can't take credit for these concepts, I believe they've been discussed here before and there are a lot of ideas in this old thread from a different forum: http://control.com/thread/997456615

 

[dmargineau]

As long as a controller will control I/O, ladder logic will be REQUIRED!

FBEs, SFCs and STLs were created and developed for process control and/or data handling/transfer applications. They substantially increase the programmers' and CPUs' workload efficiency, of course at the detriment of exponentially increasing the burden of debugging/troubleshooting issues.

 

[flux]

I find Modicon's SFC(Grafcet) implementation way easier to troubleshoot. When something get's stuck you'll know exactly which step it is by just looking at the block that is highlighted.

The blocks and transitions can be programmed with Ladder which makes it even better.

Don't know why AB did something different. Copyright issues maybe....