Hello Ignition community.
My organization is need of a test strategy for our medical device manufacturing application. In the past we have performed a rigorous, Use Case to Requirement to Test Case formula. This has proven too tedious for the level of automation we are currently at.
Envision an architecture similar to this one, taken from Inductive University as a reasonable simplification to start the discussion:
Some notes from our team on thought process:
• The problem is how to test the Ignition Gateway which manages data flow, not so much the HMI even though they’re closely related.
• Manufacturing Cells (with PLCs) talk to the Ignition Gateway, which talks to a part tracking database to find out what the state of the part is when it enters or leaves.
• The architecture is the Manufacturing Cell changes some tags and the Ignition Gateway uses that tag data to make DB calls via procs.
• It would be nice to test the Ignition Gateway directly without needing all the cells & DBs - could it be done?
• Manufacturing Cells talk to the Ignition Gateway which talks to the part tracking database to find out what the state of the part is when it enters or leaves.
• There is a need work out how to make sure the Ignition Gateway captures the right data from the Manufacturing Cell and makes the right SQL calls without requiring all the Cells and DB.
• Creating a test framework might be more trouble than it’s worth.
I have brainstormed a few strategies but am somewhat new to Ignition. Our external system integration organization is the system supplier but validating the system is incumbent upon us internally.
• Option 1: The feasibility of sending raw TCP/IP commands to the Ignition Line Control Hub would depend on whether they provide a command set, if there’s restrictive timing involved, etc. But of course there are many ways to send TCP/IP commands if the answers to the previous questions come out favorably – I’m from the LabVIEW world and it’s one option of many.
• Option 2: If not, using an RT controller (including the possibility of a single instance of the PLC used in the production system itself, with similar software) programmed in a manner to represent the rest of the production system “behind” it is the thinking for the other option (the “HIL” option)….
Expertise is greatly appreciated!