Generator Starts on Exercise, Fails Only on Real Outage After Battery Swap: Decision Tree

Why this matters

A standby generator that fires cleanly on its weekly exercise but fails to start during an actual outage, with the trouble appearing right after a battery replacement, is one of the most dangerous reliability failures because it hides during every test and only shows when the customer truly needs power. The exercise cycle starts a warm engine under no transfer load and draws relatively little from the battery; a real outage may start a cold engine that has to crank longer and harder, and the ATS sequence loads the battery differently. A wrong or weak replacement battery (insufficient cold-cranking amps, wrong group size, or a charging fault that never fully recovers it) can spin a warm engine but sag below the controller's crank-voltage threshold on a hard cold start. This tree separates a battery/charging fault from a fuel or control fault that the battery swap merely coincided with.

Symptom presentation

Scheduled exercise runs start and complete normally, week after week. During a real utility outage, the engine cranks slowly, cranks then stalls, throws an overcrank or under-voltage fault, or the controller resets mid-crank. The failures began after the battery was changed. The customer often reports the generator "worked at its test last week" and then did nothing in the storm. The controller log may show overcrank, low battery voltage during crank, or a control reset. Cold ambient temperature worsens it, because a cold engine needs more cranking torque and a cold battery delivers fewer amps.

Quick checks

Verify the replacement battery is the correct group size and meets or exceeds the manufacturer's cold-cranking-amp spec; an undersized battery is the single most common cause. Measure resting battery voltage and then crank voltage under an actual crank attempt; a healthy battery holds above the controller's minimum crank threshold (commonly the controller faults if voltage sags too low during crank). Confirm the battery charger is float-charging correctly and the battery is fully recovered, not just surface-charged. Check terminal torque and clean connections, since a new battery with a loose or corroded lug drops voltage under crank load. Load-test the battery if a tester is available. Compare a cold-start crank to a warm exercise crank to reproduce the failure mode.

Isolation tree

Branch one, reproduce the real-outage condition. Initiate a genuine utility-loss start (open the utility supply at the ATS test, or use the controller's outage-simulation), not just the exercise routine, and on a cold engine if possible. Does it now fail the way the customer described? If yes, you have reproduced it; if it starts fine warm but fails cold, lean battery/cranking. Branch two, battery capacity. Measure crank voltage. Does it sag below the controller's threshold during crank? If yes, the battery is undersized, weak, or not fully charged. Confirm CCA spec and charge state. Branch three, charging. If the battery is correct but never reaches full charge between events, the charger, its fuse, or its connection is faulty, leaving the battery marginal for the harder real-outage crank. Branch four, connection. A loose or corroded terminal on the new battery drops voltage only under the high crank current, passing the light exercise draw. Branch five, not the battery. If crank voltage holds but the engine still fails only on a real outage start, look at cold-start fueling (choke, primer, gas pressure under load) or an ATS-sequence/control timing issue, since the battery swap may be coincidental.

Confirming diagnosis

Confirm a battery cause when crank voltage sags below the controller's minimum during a cold real-outage start and an adequate, fully charged battery of correct CCA then cranks the engine through cleanly with voltage held above threshold. The reason the exercise hides this is structural: the exercise starts a warm engine that needs little cranking torque and draws modestly from the battery, so a marginal battery passes the easy test and fails the hard one. Confirm a charging cause when the correct battery passes a load test when fresh but is found undercharged before an outage event, and repairing the charger restores reliable starts; a battery that is healthy but chronically undercharged behaves exactly like a weak battery on the harder crank. Confirm a connection cause when re-torquing or cleaning the terminal eliminates the voltage sag under crank, since a loose or corroded lug drops voltage only at the high current of a cold crank and passes the light exercise draw unnoticed. Confirm a non-battery cause only when crank voltage is verified healthy through the start and the failure persists, which then points to cold-start fueling (choke, primer, gas pressure under the harder start) or an ATS-sequence timing issue, and means the battery swap was coincidental.

Remediation

Install the manufacturer-specified battery group size meeting or exceeding the required CCA, fully charge it, and verify crank voltage stays above the controller threshold on a cold start. Repair or replace a faulty charger and confirm the battery reaches and holds full float charge between events. Clean and torque all battery terminals. After any battery fix, run a real-outage simulation (not just an exercise) on a cold engine to prove the start. If the cause is fueling or control sequence rather than the battery, address that path and still verify with an outage simulation. Document the crank-voltage reading so the start margin is recorded.

References

• Generator manufacturer installation and service manual, battery group size, CCA, and charging specifications (Generac, Kohler, Cummins per model)
• NFPA 110, standard for emergency and standby power systems, starting and battery requirements
• NEC 2023 Article 700 and 702, emergency and optional standby systems
• UL 2200, stationary engine-generator assemblies
• Manufacturer controller crank-voltage and overcrank fault reference