I need to come up with a CM process for PLC code.
Currently, the system is developed using RSLogix 5000. The build product is a monolithic file that can be loaded onto a PLC for execution and edited directly in the development environment. With multiple developers, this has become a problem. They're stepping on each others changes.
As an analogy, it's as if, when doing Java development, the only wway to edit and save the source would be to load up a *.jar file into your IDE, make the change, and then save it back to the jar file. This is less than ideal.
How can I coordinate changes between multiple developers working with PLC's?
If we are talking about one big binary files, then a VCS (centralized or decentralized) is not the best tool for the job.
An external referencial (a shared disk for instance) where a batch will copy and label the current PCL state is better.
See "Tracking Software History"
To avert discontinuities in the historical record of revisions, old versions of programs must be stored.
“We take it a step further, though. Using our MDT AutoSave, we actually go out and interrogate the equipment. Overnight or at whatever frequency is specified, the software reads the programs in the PLCs and then compares that information to the last known program. The version-control software will copy the new program and store it and [then] compare it to the last one.
Launching version control is fairly simple. Required is software installation and then hardware configuration. “You would need a server and a couple of weeks of engineering and you’re good to go,” Perysyn says. However, his company uses a “shrink-wrap approach” that involves installing the software and then customization by users filling in the blanks.
That being said, when you have multiple changes from multiple developers, you need an integration environment where a first delivery can be done and validated, before pushing it to the actual server.
See also this post.