What a Load Bank Test Proves and What It Does Not
Why this matters
A load bank report with every step green gets read as "the generator is good." It proves less than that. A resistive load bank proves the engine and alternator carry rated kW at steady voltage and frequency for as long as the test runs. It says nothing about whether the unit can start the air conditioner, whether the transfer switch transfers, or whether the house wiring is right. A shop that sells the report as more than that owns the outage where the unit passed in May and stalled the compressor in July. This card draws the line. The procedure is 'Stationary Generator Annual Load Test Procedure'; whether a site needs the test is 'Load Bank Test Required vs Not Required Decision Tree'.
What a resistive load bank is
A resistive load bank is a cabinet of heating elements and a cooling fan, switched in steps, connected to the generator's output. It draws real power at a power factor of 1.0: every amp it draws is kW the engine must make. Because the load is known and adjustable, the tech can hold the generator at 25, 50, 75 and 100 percent of rating and read what it does at each step without depending on what the house happens to be running. Landing the bank's leads is done with the unit locked out of auto-start and the output dead and proven per 'The Generator Auto-Start Lockout Before Service SOP', the leads come off the same way before auto-start is restored and proven by that SOP's return-to-service steps, and the bank's hot discharge air is pointed away from people, the house and anything that burns.
That makes it the one way to put a residential standby at full rated load on demand. A house rarely draws full rated load except for a few seconds at transfer or a compressor start, so without a load bank the unit's full-load behavior is mostly unobserved.
What it proves
Held at each step long enough to stabilize, a resistive load bank test proves:
- The engine makes rated kW. If the unit cannot hold its kW rating, frequency sags and keeps sagging as the governor runs out of throttle. A worn engine, a restricted air filter, an incorrect fuel setting or a weak ignition shows up here.
- The governor holds frequency. Frequency at each step stays within the band the maker states, and returns after each step change.
- The voltage regulator holds voltage under steady resistive load, within the maker's stated regulation.
- The cooling system carries sustained heat. Twenty or thirty minutes at full load brings engine and alternator temperatures to where an outage would put them. An air-cooled unit with a blocked intake screen, or a liquid-cooled unit with a weak fan or a plugged radiator, shuts down on temperature here instead of during a storm.
- The fuel supply delivers full-load flow, on that day. Gas pressure read at the unit's test port at 100 percent shows whether the supply holds the maker's minimum. On propane that is proven only at the day's temperature, because vaporization falls as the tank gets colder; a summer pass does not certify a January morning. See 'How to Check a Propane Tank for Generator Runtime and Vaporization'.
What it does not prove
- Motor starting. A resistive bank steps up in kW at power factor 1.0. A compressor start is a burst of mostly reactive current at a low power factor, several times its running current, for a fraction of a second. The alternator's response to that, how far voltage dips and how fast it recovers, is not exercised by any resistive step. Starting arithmetic is owned by 'Motor Starting kVA and the NEC Code Letter'.
- Transfer. The load bank is usually connected at the generator's output breaker or a load bank connection point, upstream of the transfer switch. The switch never moves. Utility-fail sensing, the start delay, transfer, retransfer and cooldown are proven only by a transfer test: 'Automatic Transfer Switch Commissioning Technique' at install and 'How to Run a Simulated Outage Test on a Service Visit' afterward.
- The house wiring and the load management. Load-shed modules, branch circuits, the neutral and grounding arrangement, nothing downstream of the switch is touched.
- Reactive capability. A resistive test at power factor 1.0 does not load the alternator to its full kVA where its rating is stated at a lower power factor, as most three-phase commercial sets are. That needs a reactive load bank, below.
Worked: one test, one finding, one gap
A 22 kW air-cooled standby on propane, 240 V single phase, serving a house with the 4-ton condenser worked in 'Motor Starting kVA and the NEC Code Letter' (104 A locked rotor, 25.0 kVA starting). The values are this example's. Say the maker's sheet gives an inlet pressure range of 10 to 12 inches of water column for this unit on propane; read the real range off the real sheet.
LOAD BANK LOG (example) 22 kW LP, 240 V 1-ph
Step kW Amps Volts Hz Gas in.wc Min
0 0.0 0 241 60.2 11.0 stat 5
25% 5.5 23 240 60.1 10.6 10
50% 11.0 46 239 60.0 10.2 10
75% 16.5 69 239 59.9 9.8 10
100% 22.0 92 238 59.8 9.4 30
Engine temp stable at 100%: yes, minute 18
The electrical side passes. Amps check against kW: 22,000 W divided by 240 V is 91.7 A, logged as 92. Voltage fell from 241 at no load to 238 at full load, 3 V, about 1.2 percent of 241. Frequency fell from 60.2 Hz to 59.8 Hz, 0.4 Hz. Both are small, steady drops of the kind a healthy unit shows; compare them with the maker's stated regulation for this model rather than with this example.
The fuel side fails. Static pressure was 11.0 inches, inside the 10 to 12 range. At 75 percent it was 9.8, already 0.2 below the 10-inch minimum, and at 100 percent 9.4, 0.6 below. The engine still held 22.0 kW at 59.8 Hz, so the electrical readings alone would have passed the unit. The load bank's real finding is that the supply is marginal at full flow on a mild day. Colder weather lowers tank pressure and vaporization, so the same supply will fall further. That goes to the propane supplier and the gas fitter: regulator setting or capacity, line size, or tank size. 'How to Measure Gas Pressure at the Generator, Static and Under Load' owns that measurement.
The gap. Nothing in this log says whether the unit can start the 25.0 kVA compressor while carrying the rest of the house. The log's largest step was 22.0 kW applied as resistive load in increments, never a 25 kVA low-power-factor burst. The report goes to the customer with one line stating that motor starting and transfer were not tested and when they were last proven.
Wet stacking, and why diesel sets are exercised under load
Wet stacking is what happens to a diesel engine run for long periods at light load. Combustion temperature stays low, fuel is not fully burned, and unburned fuel, lube oil and soot collect in the exhaust system, sometimes weeping as a black oily liquid from joints. Rings and cylinders can glaze. A diesel standby that only ever exercises at no load drifts toward it, and the fix is load: run at a substantial share of rating long enough to bring exhaust temperature up and burn the deposits out.
Spark-ignited gaseous engines, which most residential standby units are, do not wet-stack in the same way, because there is no liquid fuel to pass through unburned. Light-load operation still has costs for them, but the diesel load-bank argument does not transfer directly.
NFPA 110. For emergency and legally required standby systems where it is adopted, NFPA 110, Standard for Emergency and Standby Power Systems, in the edition your authority having jurisdiction has adopted, binds the facility owner and reaches the service contractor through the facility's maintenance program. In substance, its Chapter 8 requires diesel sets to be exercised monthly under a load that either keeps exhaust temperature at the minimum the maker recommends or meets a stated minimum percentage of nameplate rating, and where monthly exercise does not meet that, requires an annual supplemental load bank test in stepped loading for a stated duration. Check the current edition for the percentages and times. Residential optional standby under NEC Article 702 is usually outside NFPA 110 entirely; see 'NFPA 110 Emergency Power Compliance' for where it applies.
Reactive load banks
A resistive-reactive load bank adds inductive steps so the generator can be loaded at a chosen power factor, commonly 0.8, the figure most commercial sets are rated at. That loads the alternator to its full kVA, heats its windings the way a real motor-heavy building does, and tests the voltage regulator against reactive load. It still does not reproduce a motor's starting transient.
On residential work, a reactive bank is rarely needed: most residential units are rated at unity or near-unity power factor and the resistive test answers the questions that matter. On commercial sets, paralleled sets, or a dispute about alternator capacity, it is the right instrument, and the report says which kind of bank was used.
Writing the report so it says only what was proven
- Name the bank type and connection point. Resistive or resistive-reactive; at the output breaker or downstream of the switch.
- Print each step's readings, not "passed." The worked log above, with the gas pressure column, is the minimum.
- List what was not tested, with the date each was last proven: motor start, transfer, retransfer, load management.
- State the ambient temperature on any propane site, because the fuel result is only valid at it.
- Route the findings. A fuel shortfall goes to the gas side, an alternator or engine shortfall to the generator side, and the customer gets one plain sentence for each.
References
- NFPA 110, Standard for Emergency and Standby Power Systems, Chapter 8, in the edition your authority having jurisdiction has adopted; check the current edition.
- NFPA 70, National Electrical Code, Article 702, as adopted by your jurisdiction.
- Generator maker's specification sheet for voltage and frequency regulation and the fuel inlet pressure range.
- See related: Load Bank Test Required vs Not Required Decision Tree; Stationary Generator Annual Load Test Procedure; How to Connect a Portable Load Bank to a Residential Standby; Motor Starting kVA and the NEC Code Letter.