Generator Just Did PM Now Wont Start On Next Exercise: Post-Service Regression Decision Tree

Why this matters

A generator that ran the cooldown after preventive maintenance, then failed to start on the next scheduled exercise cycle, is a classic post-service regression. The window between "the tech left" and "the next start attempt" hides which step caused the failure. Most cases resolve to one of four causes: something was left in the wrong state at the end of the PM, a part installed during PM is faulty, an unrelated condition was masked by the PM and surfaced now, or the start sequence is failing on a separate cause. This tree walks the right order, protects the customer relationship on a callback that feels like the tech caused it, and prevents a "make it run today" shortcut that leaves the next exercise broken again.

Section 1: Read the controller before touching anything

Pull the full event log from the generator controller covering the PM visit through the failed start attempt. You want the cooldown shutdown event, any post-cooldown status, the next exercise trigger, and the fault chain on the failed start.

Key patterns:

• Cooldown completed normally, exercise triggered, cranked for full window, no start, fault posted: a real start-cause issue.
• Cooldown completed normally, exercise triggered, no crank attempted, fault posted: a permissive or auto-mode-off issue.
• Cooldown completed normally, exercise never triggered: exercise schedule was disabled or auto mode was left off.
• Cooldown abnormal or fault during cooldown: the failure started during the PM, not after.

The pattern tells you which branch.

Section 2: Mode and permissive checks (most common cause)

A tech finishing a PM will often leave the controller in a mode that lets them safely test, then forget to restore it.

Verify:

• Controller mode: Auto, not Off and not Manual. A controller left in Off will not start on exercise or on a real outage.
• ATS in the correct state. An ATS left in Test or Bypass after a load-bank exercise will not request a start from the generator.
• Maintenance disable or service-mode lockouts cleared.
• Battery disconnect, fuel valve, and any external start permissive in the run position.

A generator in Off mode is not a generator fault. It is a procedural finding. Document and educate.

Section 3: Battery and starting circuit

PM often includes a battery test, and a marginal battery can pass a load test under tech conditions then fail on a cold exercise start two weeks later.

Verify:

• Resting battery voltage. A battery showing healthy voltage at rest but collapsing under crank load is at end of life regardless of the PM result.
• Battery cable tightness at the post and at the starter. A cable disturbed during the PM and not fully retorqued can intermittently drop crank current.
• Charger output. If the trickle charger is off, mis-set, or on a tripped circuit, the battery will be at the start of its discharge cycle when the exercise hits.
• Crank current under attempted start, compared against the manufacturer specification.

A battery that cranks slow then quits is not the same as a battery that does not crank at all. Capture which.

Section 4: Fuel and air system after PM

The PM may have changed air filter, fuel filter, or fuel system components. Any of these can fail to seal or fail to bleed properly.

Verify:

• Fuel filter installed in the correct orientation, seal seated, no air leak.
• Fuel line connections at filter or pump retorqued.
• For diesel, system bled per manufacturer procedure. A diesel that was not properly bled after fuel filter change will crank for the full window and not start.
• For natural gas or LP, gas pressure at the regulator under crank. A regulator that supplies enough pressure at idle can sag under crank demand if the gas service is borderline or the regulator was disturbed.
• Air filter installed in the correct orientation, with no missing seal.

A simple check: is the engine getting fuel during crank? If not, the fix is in the fuel system. If yes, the fix is in spark or compression or controller permissives.

Section 5: Spark, glow, and ignition

For gaseous-fueled engines, verify spark at the plugs or coil-on-plug at crank. A plug that was changed during PM may be the wrong heat range, gapped wrong, or simply faulty out of the box. A coil disturbed during PM may have a marginal connector.

For diesel engines, verify glow plug operation. A glow plug controller that failed during PM (or a glow plug that was not replaced when due) will cause a cold-condition no-start that did not show up in the warm PM window.

Section 6: Coordinate with the next exercise window and customer expectation

A clean close has the generator starting and running its exercise cycle under load successfully, with the controller back in Auto, the cause documented, and the customer told plainly what happened.

Plain framing:

• "Your generator did not start on its next scheduled exercise after the PM. Here is what we found and what we corrected, and here is what we changed in our internal PM checklist so it does not happen again."
• If the cause was procedural (mode left wrong): own it directly. Do not invent a hardware story.
• If the cause was a part installed during PM: warranty the part on the spot. Do not bill the customer for the return visit.
• If the cause was unrelated and the PM happened to coincide: explain the data clearly, with the event log timestamps, so the customer sees that the cause was independent.

Set a callback for the next exercise window to confirm successful start. A post-service regression is not closed until the next exercise runs clean.

References

• NFPA 110, Standard for Emergency and Standby Power Systems.
• NEC 2023, Article 445 (Generators).
• NEC 2023, Article 700 (Emergency Systems), Article 701 (Legally Required Standby Systems), Article 702 (Optional Standby Systems).
• NFPA 37, Standard for the Installation and Use of Stationary Combustion Engines and Gas Turbines.
• UL 2200, Standard for Stationary Engine Generator Assemblies.