Why Does a Hospital Generator Pass Its Start-Up Check but Still Risk Failing Under Load?

Direct answer: a start proves that the generator can crank, run, and produce power at that moment; it does not prove that the diesel engine can burn fuel cleanly and carry sustained emergency load. Repeated operation below roughly 30% of rated load can leave the engine too cool for complete combustion. Unburned fuel and carbon then accumulate—a condition called wet stacking—while the generator continues to look ready on short, lightly loaded checks.

That distinction matters in a hospital because standby sets are often sized for a peak emergency demand that the building rarely presents during a routine test. The exercise load may be only a small fraction of nameplate capacity. The generator can therefore pass the visible test—start, transfer, run, stop—without ever demonstrating the thermal and electrical performance the facility will need during a long outage.

Low load keeps diesel combustion too cool

Caterpillar identifies loads below 30% of rated output, sustained for extended periods, as the classic wet-stacking pattern. At light load, cylinder heat and pressure stay below the engine’s intended operating range. Some fuel does not burn completely, and unburned fuel, carbon, and oil deposits can collect in the exhaust path and around internal components.

Wet stacking is also called exhaust slobber or exhaust manifold slobber. Caterpillar describes the visible sign as a thick black liquid at exhaust-manifold, turbocharger, or pipe joints. The liquid is mainly carbon particles combined with unburned fuel and, in some cases, lubricating oil; it is not simply old oil leaking from the engine.

The 30% figure is a warning threshold, not a cliff edge. A single short run at 28% does not create the condition, and a run slightly above 30% does not guarantee healthy combustion. Duration, frequency, engine design, ambient conditions, jacket-water temperature, and the manufacturer’s minimum exhaust-gas temperature all matter. The risk comes from a repeated pattern: month after month of no-load or lightly loaded operation.

Over time, Caterpillar warns that underloading can produce deposits behind piston rings and inside cylinders and, in severe cases, cylinder-liner polishing. Those effects can lead to power loss, poorer performance, accelerated component wear, higher maintenance cost, and unplanned downtime. The start-up check may remain reassuring while that slower degradation is already underway.

A start-up check asks a smaller question

A successful start answers the starting question, not the endurance question. It can confirm that batteries, controls, fuel supply, and the starting system worked at that moment. Depending on how the exercise is configured, it may also show initial transfer-switch operation and stable voltage and frequency.

A brief or lightly loaded exercise does not necessarily test whether the generator can accept the hospital’s emergency load and keep carrying it. Sustained demand changes the question from “Will it run?” to “Will combustion, cooling, lubrication, fuel delivery, governing, and voltage regulation stay within limits while it works?”

That is why a generator can have a clean start history and still be an uncertain emergency asset. The record may prove repeated ignition without proving repeated performance at the load and duration the emergency plan assumes.

What are the warning signs of wet stacking?

The clearest field sign is black, oily slobber at an exhaust joint. Facility teams and service providers should treat that finding as a reason to review the unit’s load history, not as a cosmetic nuisance to wipe away before the next exercise.

Other useful evidence is documentary rather than visual:

• Exercise loads repeatedly below the manufacturer’s minimum or below 30% of standby nameplate kW.
• Exhaust temperatures below the engine manufacturer’s recommendation during monthly exercise.
• Service findings of carbon or deposit build-up around injectors, valves, rings, cylinders, the turbocharger, or the exhaust system.
• A pattern of short no-load runs with no periodic operation at a substantial load.
• Reduced ability to accept or hold load during testing, especially when paired with abnormal exhaust deposits or service findings.

Visible slobber alone does not prove that an engine is already damaged. Caterpillar makes that distinction clearly: slobber is a sign of possible underloading, while the more serious concern is the deposit build-up and wear that prolonged light loading can cause. Diagnosis and corrective work belong with the generator’s qualified service provider and the engine manufacturer’s instructions.

What does load bank testing reveal first?

A load bank applies controlled electrical demand when the building cannot supply enough load. The test team increases load in stages and watches the whole generator set work, rather than asking the engine to idle through another successful start.

The earliest useful information is about load acceptance and stability. As steps are applied, the team can observe whether voltage and frequency dip and recover acceptably, whether the governor holds speed, and whether the engine responds without abnormal alarms or shutdown behavior. A generator that starts cleanly can still struggle when real electrical work is demanded from it.

As load rises, the test exposes combustion behavior. Higher cylinder temperature helps burn accumulated unburned fuel and carbon deposits. Exhaust condition, exhaust temperature, smoke behavior, and any slobber can be observed while the engine is finally operating in the temperature range it was designed to reach.

At sustained higher load, the test becomes a systems check. It can reveal limits in fuel delivery, cooling and radiator performance, oil pressure, alternator output, voltage regulation, controls, and the engine’s ability to hold output over time. Those are precisely the weaknesses a no-load start may leave invisible until an outage supplies the load all at once.

Andrew Keith of Power Prove, a load-bank manufacturer, makes the vendor claim that periodic load bank testing provides a layer of assurance routine exercise runs cannot achieve. The practical point is narrower and well supported by manufacturer guidance: controlled load creates evidence about sustained performance that a start alone cannot create.

What does a load bank burn off—and what can it not repair?

A properly conducted loaded run can burn off accumulated unburned fuel and carbon deposits. Generac’s manufacturer guidance describes load-bank testing as artificially increasing load so the engine reaches operating temperature and burns accumulated build-up. Caterpillar likewise treats sustained light loading as the condition that allows deposits to form behind rings and inside cylinders.

A load bank test is not an overhaul in a box. Raising temperature may clear deposits and improve combustion, but it cannot reverse cylinder-liner polishing, worn components, fouled parts that require service, or unrelated fuel, battery, cooling, and control defects. Heavy deposits, unstable readings, leaks, alarms, or an inability to hold load are reasons to stop and escalate under the test plan—not reasons to push a suspect generator harder.

How often does NFPA 110 call for loaded exercise and load bank testing?

For diesel emergency power supply systems, NFPA 110 uses monthly loaded exercise as the baseline. The diesel set is exercised for at least 30 continuous minutes by one of two routes: loading that maintains the manufacturer’s minimum recommended exhaust-gas temperature, or operation at normal operating temperature at not less than 30% of the emergency power supply’s standby nameplate kW rating.

The percentage is calculated from the standby nameplate rating, not from the load the building happens to present. A 600 kW standby set carrying 120 kW is at 20% of nameplate, even if 120 kW is all the connected test load available. Recording both kW and percentage of nameplate makes that gap visible in the maintenance history.

When monthly exercise does not meet either NFPA 110 diesel criterion, supplemental load bank testing enters the program. The exact annual step profile depends on the NFPA 110 edition adopted and enforced for the facility. The 2025 edition and the older healthcare survey pathway summarized by ASHE’s Health Care Facilities Management use at least 50% of nameplate kW for 30 continuous minutes followed by at least 75% for one continuous hour, for 1.5 hours total.

Editions through 2022 used a different annual profile: at least 25% for 30 minutes, then 50% for 30 minutes, then 75% for 60 minutes, for two continuous hours. A facility should therefore write its procedure from the edition adopted by its authority having jurisdiction and referenced by its CMS or accreditor pathway—not from a vendor checklist or the newest edition by default.

A load bank can also be used to supplement a monthly exercise. If the building load is chronically short, supplemental load may help a monthly run reach the manufacturer’s exhaust-temperature requirement or the 30% nameplate route. That use should not be confused with the separate annual supplemental test triggered when the monthly criteria are not met.

Separate Level 1 duration testing also exists on a 36-month cycle. That test addresses the system’s required runtime class and should not be treated as proof that chronic low-load operation between those tests is harmless. The facility’s adopted standards, accreditor requirements, manufacturer manual, and authority having jurisdiction govern the complete schedule.

A practical load review before the next monthly test

Start with the last 12 months of test records. A useful review looks at the pattern, not just whether a box was checked each month.

1. Record the actual load in kW and as a percentage of standby nameplate kW. A duration without a load percentage does not show whether the diesel exercise met the thermal point of the test.
2. Record exhaust-gas temperature when the manufacturer specifies a minimum. Temperature is an alternative NFPA 110 route for diesel exercise, so the value belongs in the test record when that route is used.
3. Compare building load with supplemental load. Separate what the emergency system carried from what a load bank added; otherwise, the record can make a lightly loaded set appear better tested than it was.
4. Inspect the exhaust path and service notes. Look for slobber at joints and documented deposits, then connect those findings to the load history.
5. Confirm the governing documents. Identify the adopted NFPA 110 edition, accreditor pathway, authority-having-jurisdiction requirements, and engine manufacturer’s minimum-load and exhaust-temperature guidance.
6. Schedule corrective loaded testing when the trigger is present. Use qualified personnel, a written test plan, defined stop conditions, and the required arrangement for emergency loads to replace test loads if normal power fails during the test.
7. Document the result and the follow-up. Record load steps, duration, temperatures, pressures, voltage, frequency, alarms, corrective actions, and any service referral in the permanent maintenance record.

Running the same 15% load for longer is not a substitute for reaching the required operating condition. More time at a temperature too low for complete combustion can add runtime without answering the performance question.

Questions facility teams ask

Is 30% a hard line between safe and unsafe operation? No. Caterpillar describes wet stacking as typically associated with operation below 30% for extended periods, while NFPA 110 uses 30% of standby nameplate kW as one monthly exercise route. The manufacturer’s minimum exhaust-gas temperature is the other route. Treat 30% as a screening threshold that requires attention to load, temperature, duration, and engine-specific guidance.

Does every hospital generator automatically need an annual load bank test? No. Under NFPA 110, the annual supplemental load bank test is triggered when the diesel set’s monthly exercise does not meet the standard’s load or exhaust-temperature criteria. A facility may still schedule load bank testing more often under its own policy, the manufacturer’s instructions, an accreditor requirement, or direction from its authority having jurisdiction.

Can a load bank test fix wet stacking? It can help burn off accumulated unburned fuel and carbon and can show whether performance recovers under load. It cannot by itself repair advanced deposits, cylinder-liner polishing, worn parts, or unrelated system defects. Persistent slobber or poor loaded performance needs manufacturer-guided service, not repeated informal attempts to “run it harder.”

Does a clean exhaust mean the generator is safe from wet stacking? No single visual check proves the engine’s internal condition. A clean exhaust is only one observation. Load percentage, exhaust temperature, runtime pattern, service findings, and performance under controlled load provide the stronger evidence.

What should be escalated instead of simply retested? Inability to accept or hold the planned load, unstable voltage or frequency, abnormal coolant-temperature or oil-pressure behavior, heavy deposit discharge, leaks, alarms, protective shutdowns, or worsening exhaust slobber should be escalated to the qualified service provider under the facility’s test and impairment procedures.

The facility takeaway

A generator that starts has passed only the first test. The readiness question is whether the diesel engine regularly reaches the temperature and load needed for clean combustion—and whether records prove that it can sustain emergency output when the building finally asks for it.

For a healthcare facility team, the useful habit is simple: stop treating runtime as the headline number. Put load percentage and exhaust temperature beside it. When those numbers show chronic light loading, use controlled load bank testing to expose the problem while there is still time to correct it.

Informational only. This article does not replace NFPA 110, NFPA 99, the generator manufacturer’s manual, the facility’s emergency power procedures, or direction from the authority having jurisdiction. Testing and corrective work should be performed by qualified personnel.