This guide walks through a practical procedure for Caterpillar sets with EMCP series controllers: what to check before connecting anything, how to plan the load steps, which readings matter, and how to turn a few hours of running into a record that means something to an owner, a buyer or an insurer.
Why Cat sets need a proper load test
Standby sets spend most of their lives cold. Short, lightly loaded runs leave unburned fuel and carbon in the exhaust and on the rings, and they never stretch the cooling system, the fuel system or the alternator insulation. Problems such as a weak injector, a partly blocked radiator core or a sticking actuator often stay hidden until the set is asked for real power.
A load bank run also clears light-load deposits, confirms that protective functions behave correctly under load, and gives a baseline for later comparison. The routine service work that sits alongside it is covered in our post on Cat service intervals by hours and calendar.
Resistive or reactive load bank
A resistive bank draws power at unity power factor. It loads the engine to full kW and is the usual choice for annual testing, but it does not load the alternator to its full kVA, because a typical Cat set is rated at 0.8 power factor. A combined resistive and reactive bank can apply lagging load as well, testing the alternator, voltage regulator and excitation system closer to their real limits.
For a routine engine health check, a resistive bank is generally enough. For commissioning, acceptance after an alternator rewind, or a set that will feed large motor loads, a reactive capability is worth the extra equipment and time. Whatever bank is used, its own capacity, cooling air path and cable rating must suit the set under test.
Pre-test checks and safe connection
Before the engine starts, check oil level, coolant level and condition, fuel quantity for the full test plus margin, belts, hoses, air filter restriction and the exhaust system for leaks or loose supports. Read the controller event log and clear only what has been understood; our guide to reading EMCP alarm codes explains what the entries mean.
Connection is where most of the risk sits. The load bank should be connected at a dedicated test connection cabinet or at the generator breaker load side, with the building load isolated or the transfer switch locked in a safe position. Apply lockout and tagging, confirm phase rotation, size the temporary cables for the full test current and route them where they cannot be driven over or tripped on. Keep the bank exhaust air away from the generator intake and from people.
A stepped load plan
Start the set and let it warm up at no load until coolant temperature stabilises. Then apply load in steps. A common pattern is 25, 50, 75 and 100 percent of the test rating, holding each step long enough for temperatures to settle, often 15 to 30 minutes, and holding the top step for an hour or more. Many owners in NFPA 110 jurisdictions use the annual supplemental test pattern of 50 percent for 30 minutes followed by 75 percent for 60 minutes when monthly exercising cannot reach the required load.
Choose the test rating with care. A standby rating is time-limited by definition, so extended running is normally done at the prime rating or a defined percentage of nameplate. Our page on prime and standby duty language for Cat-class sets explains the difference.
| Step | Load | Typical hold | Main purpose |
|---|---|---|---|
| 1 | No load | Until warm | Stable oil pressure and coolant temperature |
| 2 | 25 percent | 15 to 30 min | Check voltage and frequency regulation |
| 3 | 50 percent | 15 to 30 min | Exhaust temperatures begin to separate weak cylinders |
| 4 | 75 percent | 30 to 60 min | Cooling system and fuel system under sustained duty |
| 5 | 100 percent of test rating | 60 min or more | Full output, alternator heating, governor margin |
| 6 | No load | About 5 min | Cooldown before stop |
Readings to log at each step
Record readings at the start and end of each step, and every 15 minutes during long holds. From the EMCP display or service software: kW, kVA, power factor, voltage on all three phases, current on all three phases, frequency, engine speed, oil pressure, coolant temperature and, where available, fuel rate and boost pressure. From the engine itself: exhaust temperatures on each bank or each cylinder where thermocouples are fitted, charge air temperature and visible smoke.
Also note ambient temperature, altitude and any derate active on the controller. A set that loses output on a hot afternoon may be behaving exactly as its derating curve says, which is a different conclusion from a set with a fuel or air problem.
Block load and recovery checks
Stepped loading proves endurance. It does not prove how the set reacts when a building drops its whole load onto it in one go. Where the bank allows it, apply a block step, for example from zero to 50 percent or from 50 to 100 percent, and record the frequency and voltage dip and the time to recover. Repeat for load rejection.
Compare the results against the performance class the set was bought for, often expressed against ISO 8528-5, and against the needs of the site. Large dips may point to governor tuning, a tired actuator or turbocharger lag rather than to the set being undersized. The load acceptance page on this site covers how the same behaviour is judged when a site later adds gas turbines.
Reading the results and writing the record
A healthy set holds voltage and frequency inside their bands at every step, keeps coolant temperature stable below alarm levels, shows exhaust temperatures that rise evenly across cylinders, and carries full test load without smoke or derate. Warning signs include one cylinder running much cooler or hotter than its neighbours, coolant temperature that keeps climbing at full load, falling oil pressure as the oil heats, and frequency that sags under sustained load.
The written record should list the set serial number, hour meter reading before and after, load bank details, ambient conditions, the full table of readings, any alarms, and the technician’s conclusions with recommended actions. For a used set being considered for purchase, that record carries real weight alongside oil analysis and inspection findings, and our asset refurbishment team can test a set before it ships.
About the authors
This guide comes from the USP&E commissioning engineers and field technicians who load test Cat-class generator sets on mining, utility, industrial and standby sites. USP&E is an EPC and O&M contractor, not a broker, with 150+ power projects in 45+ countries, 25,000 MW contracted since 2002 and 24+ years in the field. Testing forms part of the routine set out on our operations and maintenance page.
Offices: Colorado Springs (1755 Telstar Dr, Suite 300), Knoxville (800 S Gay St, Suite 700), London (162-168 Regent Street), Johannesburg (22 Kildoon Rd, Bryanston), Cape Town (31 Brickfield Road, Woodstock) and China (No. 1 Ningzhou Road, Wuxi). Regions: USA | AFRICA | UK | UAE | CHINA. ISO 9001 and ISO 45001 certified.
Cat load bank test questions
How long does a load bank test on a Cat generator take?
A typical stepped test takes two to four hours including warm-up and cooldown. Commissioning or acceptance tests with block load steps can take longer.
How often should a Cat standby generator be load bank tested?
Many owners test annually, and more often where monthly exercising cannot reach a meaningful load. Local codes, such as NFPA 110 where it applies, may set minimum requirements.
Can I load test a Cat generator using the building load?
Yes, if the building load is large and steady enough, but it is rarely controllable. A load bank allows exact steps and full load without disturbing the facility.
Do I need a reactive load bank?
For routine engine checks a resistive bank is usually enough. A reactive bank is worth using for commissioning, after alternator repairs, or where the set feeds large motors.
What readings should be recorded during the test?
Electrical values on all three phases, frequency, oil pressure, coolant temperature, exhaust temperatures, fuel rate and boost where available, plus ambient conditions and any alarms.
Does load bank testing fix wet stacking?
A sustained high-load run usually burns off early light-load deposits. Heavy or long-standing deposits may need repeated runs or mechanical inspection.
Plan a load bank test
Tell us the Cat set model, rating, site and what you want the test to prove, and a USP&E technician will propose a test plan and record format.
If a document set is easier than a conversation, send it to info@uspeglobal.com or upload a summary through the test planning form; to talk it over, dial +27 10 822 2324. Tested sets ready to ship are listed in our diesel generator inventory.
