A Standby Generator That Starts Is Not One That Will Carry the Load

Most standby generators are tested the same way: started once a month, run for fifteen minutes off load, logged as tested, shut down. The engine starts, the log fills up, and everybody is reassured. What has been demonstrated is that the battery, the starter and the fuel solenoid work. Nothing has been demonstrated about whether the set will carry the building.

What a no-load run hides

A diesel engine running with almost no load never reaches working temperature or cylinder pressure. Unburnt fuel and soot accumulate in the cylinders, around the rings and in the exhaust — the condition usually called wet stacking. It shows as a wet exhaust stack and black residue, and it does more than look bad: it glazes bores, fouls injectors and progressively reduces the output the engine can actually deliver. The monthly test intended to prove readiness is, slowly, removing it.

A no-load run also leaves the parts most likely to fail completely untested: the alternator under real excitation, the governor holding frequency against a step change, the cooling system at full heat rejection, and the transfer switch actually transferring.

What a load test establishes

• It holds voltage and frequency when load is applied in steps, which is how load arrives in reality — not gradually.
• Cooling is adequate at full output, in local ambient temperature, which is where sets sized from a catalogue often disappoint.
• Fuel delivery sustains full load rather than only idle, which exposes a partly blocked filter or a failing lift pump.
• The transfer switch operates under load and returns correctly when mains comes back.
• The set still makes its rated output. Capacity falls with age, and a set that has wet-stacked for years may be well below its plate.

A practical regime

Keep the monthly start — it catches batteries, which remain the most common cause of a failure to start. But add a periodic load test that takes the set to a substantial proportion of its rating for long enough to reach temperature, using building load where that is acceptable or a load bank where it is not. Record output, temperatures and fuel pressure, so capacity can be compared year on year rather than assumed.

The question is not whether your generator starts. It is whether it will still be running, at full output, an hour into the outage.