An operator doesn't want "Fault 12" — they want which sensor, which step, why. ProDiag supervises your program with no alarm code to write, and its criteria analysis names the first-up cause. Pair it with WinCC alarm classes and acknowledgement and the machine tells its own story. Break the cell below and read what it says.
Two supervisions watch this cell. A reaction (action) supervision says Clamp_Closed must go TRUE within 2.0 s of the clamp command; an operand supervision says Tank_Level must stay below 95 %. Jam the clamp or overfill the tank and watch the alarm appear in the WinCC view — then read the criteria analysis, which names the exact condition that failed.
That criteria-analysis line is the whole point of ProDiag: instead of a bare fault number, it reports which operand in the supervised step was wrong and when — the first-up cause. You configure supervisions on the program blocks (no alarm-handling code), assign each an alarm class (Warning / Alarm / Fault — which sets colour, priority and whether an acknowledgement is required), and WinCC and the CPU's web server display them for free.
🔎 Supervisions. Attach monitoring to your program: operand (a tag must hold a value), action/reaction (an effect must follow a command within a time), and interlock checks — configured in the editor, generating the alarm automatically.
🎯 Criteria analysis. When a supervision trips, ProDiag captures the initial cause — the first operand that failed and its state — so the message points at the real fault, not a downstream symptom.
🏷️ Alarm classes. Warning, Alarm, Fault (and your own) set colour, priority, and whether the alarm must be acknowledged. One classification drives the look and the lifecycle everywhere it shows.
🖥️ One source, many screens. The alarms live in the CPU; WinCC Comfort/Unified panels, the TIA alarm view, and the S7-1500 built-in web server all subscribe to the same set — no duplicate alarm lists to maintain.
The shift is from writing diagnostics (a wall of IF fault THEN set message) to configuring them. ProDiag generates the supervision logic and the message text; you spend your effort deciding what's worth an alarm and what class it belongs to — the same alarm-rationalisation discipline good machines live by.
A step times out because a clamp never reached its closed sensor. What does ProDiag's criteria analysis add over a plain "Step 4 timeout" message?