Evidence worth collecting before the condition changes
- Decision that runtime information must support
- Existing PLC, VFD, contactor, counter, meter, or machine-state signals
- Shift or observation window
- Known difference between powered, running, loaded, and productive states
- How the result will be verified against actual operation
Questions that improve the diagnosis
- Do you need powered time, motion/run time, loaded time, productive time, cycles, downtime, runtime plus energy, or OEE inputs?
- Does an existing control signal already represent that state reliably?
- What false positives or false negatives can the chosen signal create?
- Who will verify the measured state against the physical machine?
What not to conclude too early
- Calling mains-on time 'runtime' without defining the state
- Adding a sensor before checking existing signals
- Using current alone as proof of useful production without validation