How to Verify a Generator Auto-Start on Utility Loss
A standby generator that has never been proven against a real outage is a hope, not a backup. The auto-start chain - utility sensing, start delay, cranking, voltage buildup, transfer, retransfer, cooldown - involves the transfer switch, the engine controller, and the battery all cooperating, and any link can quietly rot between outages. This page walks through verifying the sequence as designed: not inventing settings, but proving that the documented sequence actually happens, on time, with the evidence logged. It is written for the technician or operator assigned the test at a pump station, compressor site, or facility with a standby set.
Verify Generator Auto-Start on Utility Loss in one line: To verify a generator auto-start on utility loss, first obtain the site's documented sequence and settings - the start delay, transfer delay, retransfer delay, and cooldown time configured in the ATS and engine controller - then run a planned test in which utility power is removed by the approved method for that site, each stage is timed against the documented values, and the transfer and eventual retransfer are observed at the loads. The test is coordinated with operations because loads will see the outage on an open-transition switch, the switching itself is performed by qualified personnel under the site's electrical program, and the result is a written log of what happened at each stage compared with what was supposed to happen.
Know the Designed Sequence Before You Test It
Verification means comparing behavior against a specification, so the first step is paperwork: the ATS settings sheet and the engine controller's configured delays. The canonical sequence runs: utility voltage falls out of tolerance; the ATS or controller waits a deliberate start delay so a momentary blip does not start the engine; the engine cranks and fires; the generator builds to acceptable voltage and frequency; a transfer delay elapses; the ATS transfers the load. When utility returns, the mirror image: a retransfer delay long enough to trust the utility is really back, retransfer of the load, then an unloaded cooldown run before shutdown. Every one of those numbers was chosen by someone; your job is to observe whether reality matches them, and to flag mismatches rather than retune anything on the spot - setting changes go back through whoever owns the design.
The test plan needs operational honesty about what will happen: with a standard open-transition switch, every load on the ATS sees the outage during the test exactly as it would in a real one. That means coordinating with operations on timing, knowing which processes will trip and how they recover, and having the site's procedure for the approved way to simulate the utility loss - many transfer switches provide a test function that simulates loss of source, and where the method is physically opening an upstream device, that switching is qualified-person work under the site's electrical safety rules.
Running the Test and Timing Each Stage
Run the sequence with a watch and a written log. Record the moment utility is removed; the elapsed time until crank begins, which should match the start delay; crank duration and successful fire; the time for voltage and frequency to stabilize; and the moment the ATS transfers, against its configured delay. At the loads, confirm what actually came back: the pumps restarted per their own restart logic, the control panel rode through on its UPS or rebooted as designed, the comms gateway recovered. A sequence can be electrically perfect and operationally broken - generator running, loads transferred, but a PLC waiting for a manual reset nobody planned for - and the test exists to find exactly that.
Then finish the test: restore utility and verify the return path, which is the half that ad-hoc tests skip. The retransfer delay should hold the load on generator while the utility proves itself, retransfer should land cleanly, and the engine should run its cooldown unloaded and stop. Afterward, walk the site back to normal: ATS back in auto, engine controller in auto, battery charger charging, block heater and fuel valves as found. A surprising fraction of real-outage failures are artifacts of the previous test - a controller left in manual, a breaker left open - which makes the post-test walkdown part of the verification, not housekeeping.
What to Log, and the Failures That Hide Between Tests
The written record is the deliverable: each stage's expected and observed timing, what started and what did not, and any deviation flagged to the owner of the settings. If the site's telemetry carries generator and ATS status - run signal, ATS position, fail-to-start, low battery - verify during the test that each status actually arrived in the SCADA system with the right state and timestamp. An unmanned site's outage response is only as good as those points, and a platform such as Merobix trending engine runs, transfer events, and battery voltage turns each monthly test and each real outage into a comparable, archived record.
Between tests, the failure statistics have a clear headliner: the start battery. Batteries age, chargers fail quietly, and a set that exercised beautifully last quarter can fail to crank on a cold night; battery voltage during crank is one of the most diagnostic values a telemetry system can record. The other repeat offenders are fuel - stale, gelled, or simply low - coolant heaters that quit, and controllers left out of auto. Life-safety installations governed by NFPA 110 carry their own mandated testing regimes; industrial standby sets follow the site's maintenance program, and the honest measure of any program is whether the generator's last proven, timed, logged sequence is recent enough to bet the site on.
Frequently Asked Questions
Why does the generator wait before starting when utility fails?
The start delay exists to filter momentary blips: reclosers and switching events cause brief dips that restore themselves, and starting the engine for each would accumulate pointless starts and wear. The delay is a designed value in the ATS or engine controller, chosen for that site's tolerance, and the verification test confirms the observed wait matches the documented one rather than judging what the number should be.
What if the generator starts but the load never transfers?
That splits the failure cleanly: the engine side worked and the transfer side did not. The usual suspects are the ATS not sensing generator voltage as acceptable, a failed transfer mechanism or control circuit, or the switch left out of auto after previous work. It is a finding to document and hand to qualified personnel - the test's value is precisely that it localizes the failure to the transfer path while everyone is watching, instead of during a storm.
How often should a generator auto-start be verified?
Life-safety and code-governed installations follow the testing intervals their codes mandate, NFPA 110 being the reference for emergency power systems. Industrial standby sets follow the site's own maintenance program, and the practical guidance is that exercise runs without a verified transfer do not prove the sequence: periodically the full chain - sensing, start, transfer, retransfer, cooldown - should be observed and logged end to end, because the transfer path is where untested failures accumulate.
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.