The generator starts. The automatic transfer switch throws. The UPS rides through the bump without a single blip on the monitoring screen. The chiller plant, sequenced to come online in a staggered ramp, follows its programmed logic. Every piece of equipment performs exactly as specified. The test is declared a success. The facility is signed off.
Six weeks later, the utility flickers—nothing dramatic, just a momentary sag. The UPS catches it. The generator starts. The transfer switch begins its sequence. And then, in a cascade that no single piece of equipment could have predicted, the cooling plant drops offline.
What happened? The chiller’s control transformer was inadvertently fed from a panel that sits upstream of the UPS—a common installation oversight. When the utility sagged, the control power dropped out for just long enough to reset the chiller’s PLC. The PLC lost its state and didn’t automatically restart when power returned because the sequencer was waiting for a “ready” signal that never came. The BMS logged an event, but the alarm priority was set too low for the NOC to notice immediately. By the time operators responded, the hot aisle had already breached the threshold.
The equipment passed every startup test. The integrated system did not.
This scenario is not hypothetical. It reflects failure modes documented in post-incident reviews across the industry. And it exposes a fundamental gap in how many projects approach the final verification phase.
Why Vendor Start-Up Is Not Data Centre Commissioning
Manufacturer start-up is essential work. The OEM engineers validate their equipment against factory specifications. They run load banks, check voltage regulation, verify internal transfer times. The chiller vendor starts their machines, confirms refrigerant pressures, logs operating temperatures. The generator supplier exercises their unit under load and certifies the automatic transfer switch.
All of this is necessary. None of it is sufficient.
The problem is that vendor start-up verifies equipment in isolation. The manufacturer’s scope terminates at their output terminals. They do not test the communication handshake between the UPS and the BMS when the UPS sends an alarm indicating it has transferred to bypass. They do not verify that the BMS interprets that signal correctly and adjusts its sequencing logic for the cooling plant accordingly. They do not simulate what happens when the generator’s automatic voltage regulator interacts with the UPS’s input filters under a specific load condition.
These are not hypothetical edge cases. A significant portion of electrical faults in the first five years of operation trace back to termination and cabling issues—the connections between equipment, not the equipment itself. The device passes. The installation fails.
Integrated Systems Testing (IST): Who Writes the Script Matters
The objection is predictable: “But we do run integrated systems testing. We simulate transfer events. We test the sequences.”
The question is not whether IST is performed. The question is who develops the test procedures, who witnesses the execution, and who has the authority to demand that a test be re-run when the conditions aren’t right.
In standard commissioning practice, the commissioning authority writes the Functional Performance Test procedures and the Integrated Systems Test scripts. The contractor executes the tests. The CxA witnesses, documents, and verifies. This separation of responsibilities is critical because the person writing the script determines what gets tested—and what gets overlooked.
Here is where the illusion takes hold. A project team can run through an IST script, check the boxes, and declare success—while missing the critical details that will surface in production. The UPS transfers. The generator starts. The chillers sequence. On the surface, everything works.
But did the BMS actually receive the correct alarm from the UPS? Did it log the event with the proper timestamp and priority? Did the notification engine trigger the right alert to the operations team? Was the chiller’s control transformer supplied from a source that remained stable throughout the transfer? Did the automatic transfer switch’s setpoints coordinate with the UPS’s bypass logic?
These are not questions that a vendor startup script answers. They are not questions that a self-witnessed IST answers either—unless the person writing the script and witnessing the test has the technical depth to recognise what isn’t being tested and the contractual authority to stop the test and demand corrections.
What an Independent Commissioning Authority (CxA) Actually Verifies
A commissioning authority serves as the owner’s independent technical authority. The CxA reviews submittals of commissioned equipment and controls, ensures systems are properly installed prior to startup, witnesses vendor startup of critical equipment, and develops the functional and systems testing procedures that define how integration will be verified.
Consider the difference in practice.
A vendor startup engineer runs through a checklist. They verify that the UPS transfers from normal to bypass within the specified time. They note the result. They move on.
A CxA witnessing that same test does not just note the result. They ask: What was the BMS doing during that transfer? Did it register the event? Did it maintain communication with the chiller plant? What happens if we force a BMS communications failure during the transfer? What happens if the transfer occurs while the chiller is in the middle of a start sequence?
The CxA designs the IST script to find problems. They deliberately force failures. They test the permissives and inhibit logic that prevent conflicting operations. They verify that electrical transfers coordinate with cooling interlocks. They prove that the system fails safely—not just that it works under ideal conditions.
This distinction matters because the CxA is the only party in the room whose sole performance metric is the facility operating as an integrated whole. The vendor’s metric is their equipment passing. The contractor’s metric is the project finishing on schedule. The CxA’s metric is the system working—all of it, together, under every condition it might encounter.
The Cost of Skipping Independent Commissioning Verification
The financial implications are stark. Industry data shows that three out of four unplanned outages are preventable with proper testing and maintenance procedures. Yet when a facility is signed off based on vendor start-up and self-witnessed IST alone, the owner is accepting a verification that stops at the equipment terminals. The integration gaps—the communication handshakes, the control logic interactions, the coordination of setpoints—remain unverified until they fail in production.
The cost of that failure is measured in downtime. A significant portion of severe outages now result in direct financial losses exceeding $100,000, with a notable percentage crossing the $1 million mark. Project data reveals that most delays and budget overruns cluster around the turnover and integration phase—the very period when verification is most critical and most often rushed.
The CxA does not prevent all failures. No verification process can. But the CxA shifts the probability. By witnessing vendor startup, the CxA ensures that the equipment is actually ready before integration testing begins. By writing and witnessing the IST scripts, the CxA ensures that the integration is actually tested—not just run through a script. By having the contractual authority to stop a test and demand corrections, the CxA ensures that problems are found and fixed before the facility goes live, not after.
The Bottom Line
The manufacturers will always do their job. They will start their equipment, run their tests, and certify their products. That is essential work, and it should never be replaced.
But vendor start-up is not commissioning. Self-witnessed IST is not independent verification. The equipment can pass every test and the system can still fail—because the tests never asked the right questions.
The CxA asks the right questions. Not because they know more about the equipment than the manufacturer—they don’t. But because they know more about the system than any single vendor. They understand that a UPS and a chiller don’t care about each other’s factory test reports. They only care about whether they can work together when everything goes wrong.
The question is not whether to run IST. The question is who writes the scripts, who witnesses the execution, and who has the authority to say “that’s not good enough” when the test passes on paper but fails in reality.
Independent Commissioning Authority Services
At Mission Critical Engineers (MCE), our CxA services provide the independent verification that vendor start-up and self-witnessed IST cannot deliver. We write the test scripts, witness the execution, and verify the results—with no commercial interest in the equipment passing.
Our commissioning authority services include:
- Independent review of submittals and design documentation
- Witnessing and verification of vendor start-up
- Development and execution of Functional Performance Test procedures
- Development and witnessing of Integrated Systems Testing
- Documentation of results and management of issue resolution
- Final handover verification and operations team training
Choose the witness who has nothing to gain from a pass and everything to lose from a failure.
Choose independence. Verify the integration. Commission the gaps.