How to Build a SCADA Point Commissioning Checklist
Commissioning a point means proving that one signal travels correctly from the field all the way to where it is used, and doing it the same thorough way for every point is what a checklist guarantees. Without one, some points get a careful checkout and others get a glance, and the glanced ones carry the errors into production. This guide is for the engineer building the per-point commissioning checklist, so every point earns the same proof before it goes live.
Build a Point Commissioning Checklist in one line: To build a point commissioning checklist, list the checks every point must pass before go-live: correct device mapping, correct scaling and units, good quality when healthy and bad quality on failure, working alarms at the right limits, and correct recording in history. Make it identical for every point so none gets skipped, run it with a real or injected signal end to end, and record the pass so the commissioning is documented rather than assumed.
List Every Check a Point Must Pass
A good commissioning checklist covers a point from the field to every place its value is used, and the items follow that path. Confirm the mapping points at the right device and register, so the point is actually reading what it claims. Confirm the scaling and units, so it reads the right magnitude, not just a moving number. Confirm quality behaves - good when the device is healthy, bad when it fails. Confirm any alarms trip and clear at the right limits. Confirm the value records correctly in history. Each is a distinct failure that a glance would miss.
Order the checks along the signal path so the checklist reads like the journey of the value. Field signal to device mapping, mapping to scaled value, scaled value to display and quality, value to alarm, value to history - working in that order means each check builds on the one before, and a failure localizes to the step where the path breaks. The ordered checklist below captures the standard sequence:
- Mapping points at the correct device and register.
- Scaling and units read the correct value at both range endpoints.
- Quality is good when healthy and bad when the device is cut.
- Alarms trip, annunciate, and clear at the configured limits.
- The value records correctly in the historian.
- The commissioning result is recorded against the point.
Run It End to End With a Real Signal
Commission with an actual signal, injected or produced, not by reading the configuration. Drive the field end to known values and confirm each checklist item against what actually appears downstream, because the whole point of commissioning is proving the real path works, which desk-checking cannot do. A point that looks perfectly configured can still be mapped to the wrong register or scaled to the wrong range, and only a real signal traveling the path reveals it.
Check the scaling at the range endpoints, injecting the equivalent of the low and high signal and confirming the displayed engineering value matches, which proves the engineering unit scaling is right at both ends. For a nonlinear point add a midpoint. This endpoint check catches an offset or span error that a single mid-range reading would let slip through, and it is one of the highest-value items on the whole checklist.
Exercise the failure side of the point, not just the healthy reading. Cut comms to the device and confirm the point goes to bad quality and, if it has alarms, does not fire spuriously on the frozen value - which proves the point participates in your quality-flag strategy. A point that reads correctly when everything is fine but freezes and lies when its device dies is only half commissioned, and the failure half is the one that matters in production.
Make It Identical for Every Point and Record It
Run the identical checklist for every point, because consistency is the entire value of having one. The moment some points get the full checkout and others get a glance, the glanced ones become the source of the errors that surface after go-live, and you cannot tell which points were actually proven. A checklist applied uniformly means every point carries the same proof, and there is no untested remainder hiding among the tested majority.
Record the result against each point so commissioning is documented, not remembered. A checklist that is worked but not recorded leaves you unable to answer later whether a given point was actually commissioned or just assumed, which is exactly the question a post-go-live problem raises. The recorded result also feeds the broader sign-off, so the point-level commissioning rolls up into the system-level configuration validation as documented evidence rather than a vague assurance.
Template the checklist so it scales to a large point count without becoming a burden. The same checklist applies to every point of a given type, so it can be a standard form rather than a fresh invention per point, and where points come from templated assets the commissioning can follow the template too. This keeps thorough per-point commissioning practical even for thousands of points, in the same spirit that organizing tags into UDT templates keeps configuration itself manageable at scale.
Verifying the Checklist Did Its Job
Confirm coverage by reconciling the recorded commissioning results against the full point list. Every point should have a recorded pass, and any point without one is a point that was skipped and is about to go live untested. This reconciliation is the check on the checklist itself - it is how you know the commissioning covered everything rather than most things, and it turns up the device nobody got to before that device's errors turn up in production.
Spot-check a sample of commissioned points independently to confirm the checklist was actually run honestly, not just marked complete. Re-verify a few points across different types and devices against their recorded results, because a checklist that was checked off without being truly exercised provides false confidence. A small independent audit of the commissioning records is cheap insurance that the documented passes reflect real tests, and it belongs alongside the final go-live sign-off.
Common Mistakes to Avoid
The defining mistake is inconsistent commissioning - some points thoroughly checked and others glanced at - so the glanced points carry hidden errors and you cannot tell which points were really proven. Run the identical checklist for every point. The second mistake is desk-checking the configuration instead of driving a real signal end to end, which misses mapping and scaling errors that only appear when a value actually flows.
The third mistake is commissioning only the healthy reading and never the failure behavior, so a point that freezes and lies when its device dies passes as good. Exercise the bad-quality path for every point. The fourth is working the checklist without recording the result, leaving you unable to prove later which points were commissioned - record every pass and reconcile the records against the full point list so no point ships untested.
Frequently Asked Questions
What should a SCADA point commissioning checklist include?
Every check a point must pass from field to use: correct device and register mapping, correct scaling and units verified at the range endpoints, quality that is good when healthy and bad when the device is cut, alarms that trip and clear at the right limits, and correct recording in history. Order the items along the signal path so a failure localizes to the step that broke, and record the result against each point so commissioning is documented.
Why commission a point with a real signal instead of checking the configuration?
Because a point can look perfectly configured yet be mapped to the wrong register or scaled to the wrong range, and only a real signal traveling the whole path reveals it. Driving the field end to known values and confirming what appears downstream proves the actual path works, which desk-checking cannot. Check the range endpoints to catch offset and span errors, and cut comms to confirm the point fails to bad quality rather than freezing.
How do I keep point commissioning consistent across thousands of points?
Use one identical checklist for every point, templated as a standard form rather than reinvented per point, and where points come from templated assets let the commissioning follow the template too. Consistency is the whole value - the moment some points get a full checkout and others a glance, the glanced ones carry the errors. Record every result and reconcile against the full point list so no point ships untested and none hides among the tested majority.
Automation services
Need help turning this into a working system?
Merobix integrates SCADA, programs Allen-Bradley and Siemens PLCs, and designs and fabricates industrial control panels.
Meeting requests are reviewed before confirmation.