"4-hour load bank test" is one of those phrases that arrives from outside the shop. A hospital forwards an accreditation question. A spec section calls for it by name. An AHJ letter asks when the last one was run. Whoever it comes from, the shop now has to price a half-day test, plan it so nothing gets mobilized twice, and produce a record that satisfies the person who asked.
This guide covers what the term means, where the number comes from, and the planning that separates a clean test day from an expensive one.
What a 4-hour load bank test is
It is an extended-duration load bank test: the generator carries a substantial, controlled artificial load for hours, not minutes, while a technician records readings throughout. The load bank supplies the load the building cannot, in defined steps, and turns the output into heat.
Duration is the whole point. A short periodic test proves the set starts, transfers, and makes power. Hours under load prove something a short run cannot: that the machine holds together at full operating temperature over time. Cooling systems reach thermal saturation and either hold or creep. Fuel supply, day tank, and transfer pumps see sustained draw. Voltage regulation and frequency get watched for drift instead of sampled once. Weak points that hide inside a 30-minute window (a marginal radiator, a clogged fuel pickup, a thermostat on its way out) get hours to reveal themselves with a load bank connected instead of during a real outage.
Where the four hours comes from
The phrase is a term of art. Long-duration test provisions appear in emergency-power compliance frameworks; accreditation FAQs for healthcare facilities discuss the 4-hour test by name; and job specifications write it into contract scope. That is why the number arrives at the shop already attached to an obligation.
What the phrase does not tell you is the enforceable detail for a given site. Which systems owe a long-duration test, on what cycle, at what loading, with what pass criteria: those come from the edition of the governing standard your jurisdiction adopted, from the accreditor where one applies, from the AHJ, and from the specification in front of you. Editions differ, occupancies differ, and accreditors layer their own expectations on top.
The honest framing
Planning the visit
Most of what goes wrong in a long-duration test goes wrong before the engine starts. The planning list is where the margin is:
- Step profile first, hardware second. Write down the load steps, target kW, and hold times from the governing sources. Every other decision follows from that sheet.
- Load bank sizing. The bank must hold the highest step with headroom at the generator's voltage and phase. A bank running at its own limit for hours is a second machine being stress-tested in the middle of your proof.
- Fuel math. Hours at substantial load is sustained draw. Confirm on-site fuel covers the run plus the reserve the facility must keep, and schedule a delivery ahead of the test where it will not.
- Cabling and placement. Cable runs sized and routed, connections torqued, and the bank placed where its exhaust heat and noise have somewhere to go, all confirmed on a site walk rather than on test morning.
- Staffing for the duration. Someone qualified stays on readings for the entire run. Trends are the product: coolant temperature creeping after hour two matters more than any single value.
- The facility's risk window. Agree on what happens if utility power is lost mid-test, who can call an abort, and how the building's load is protected while the test owns the generator.
- Double-counting, settled in writing. If the visit is meant to satisfy more than one testing obligation, confirm that with the AHJ and any accreditor beforehand, and file the confirmation with the record.
Running it
The run itself is disciplined and boring, which is the goal. Load comes up in the planned steps. Readings (kW, percent of nameplate, voltage, amperage, frequency, oil pressure, coolant temperature, exhaust observations) get recorded at a consistent interval and at every step change. The technician watches trends across the hours, since the failures worth finding announce themselves as drift long before they trip a shutdown. At the end comes a stepped unload and a cooldown run, the unit goes back to AUTO, and every alarm or anomaly from the run gets written up with its time and load step.
The record is what survives
Six months later, nobody remembers the test; the record speaks for it. A defensible long-duration test record captures the unit's identity and nameplate, who ran the test and who witnessed it, the planned profile beside what was held in fact, the readings at interval, observations, deficiencies, and sign-offs. The free load bank test log at the end of this page (printable PDF and Excel) is that format, ready to use.
One habit worth building: record the profile you were asked to run and its source (spec section, accreditor letter, AHJ correspondence) on the test record itself. When the next auditor asks why the test looked the way it did, the answer is stapled to the evidence.
Where PowerOps fits
PowerOps does not decide your test duration; the governing sources do. What it does is make the proof durable: the load bank result becomes a structured record attached to the specific generator, committed by a named, authenticated technician with a server-stamped time, and held immutable after sign-off. The unit's identity, site, and history auto-flow into the record, so the technician spends the four hours watching the machine instead of re-typing what the office already knows.