Standby power: reading the set dossier

A standby generating set is only as good as the paperwork behind it. The engine make sets which manual and which code list apply, the controller states and event log tell you what the set did and why, and the load test, fuel estimate and maintenance schedule are what keep the transfer honest. Read those three layers in that order and a building can decide whether its backup source is verifiable or merely present.

Start with the engine make, not the nameplate rating

The first page of any set dossier is the engine. A Cummins, Perkins, MTU, Yanmar, Kubota, Isuzu, Deutz, Doosan, Iveco or Mitsubishi engine each carries its own machine code, serial number and documentary version, and those three items decide which parts list, torque table and fault code chart are valid. A set rated at the same kVA with a different engine is a different machine for maintenance purposes.

Record the engine make, the machine code and the serial number exactly as stamped, then note the documentary version the manual belongs to. When a workshop orders a filter or a sensor, the counter staff work from those numbers, not from the enclosure label. Where a building keeps several sets, a single sheet per engine make prevents the common error of applying one maker’s valve clearance to another maker’s head.

For readers who work in Vietnamese, the technical guide at reading a generator set dossier walks through the same engine-by-engine document reading for Cummins, Perkins, MTU, Yanmar, Kubota, Isuzu, Deutz, Doosan, Iveco and Mitsubishi, with machine code, serial number and documentary version treated as the identifying trio.

What the controller states, alarms and event logs actually say

The controller is the set’s own record keeper. Deep Sea and ComAp units are read by state rather than by guesswork: stopped, preheat, cranking, running, cooling, fault. Each state has a defined entry and exit condition, and the event log timestamps every transition. An alarm is not a verdict, it is an entry with a time against it.

Read the log backwards from the last event. A low oil pressure alarm that appears two seconds after cranking is a different fault from one that appears twenty minutes into a run. A high coolant temperature alarm that follows a load step points at cooling or at the load, not at the sensor. Overcrank, underspeed, overfrequency and fail-to-synchronise entries each narrow the field before a technician opens a panel.

The transfer side has its own record. An ATS cabinet coordinates with the incoming ACB or moulded case breaker, and the timing between mains failure, engine start, breaker open and breaker close is set, not accidental. If the event log shows the generator breaker closing before the engine reached rated speed, the transfer timing needs review, and the log is the evidence.

How is the automatic transfer coordinated with the breakers?

Automatic transfer is a sequence with interlocks, and the dossier should show the settings. The mains breaker opens on confirmed loss, the engine cranks and reaches rated voltage and frequency, the generator breaker closes, and the mains breaker is mechanically and electrically locked out from closing onto a live generator. The ACB or the smaller breaker in the ATS carries the short-circuit and overload protection, and its settings must match the set’s fault level.

Two numbers matter at this point: the time delay on start and the time delay on transfer. A start delay that is too short produces nuisance starts on momentary dips. A transfer delay that is too short closes the generator breaker onto a load still energised by a decaying motor. Both are adjustable, both should be written down, and both should be checked after any controller replacement.

Return transfer deserves the same attention. The mains must be stable for a set period before the load returns, and the generator should run on cool-down before it stops. A log that shows repeated short transfers between mains and generator is a sign that the stability window is set too tight for the supply.

What does a load test prove, and what does it not?

A load test proves that the set can carry a known load for a known time at rated voltage and frequency. It does not prove that the building’s essential load list is correct, and it does not prove that the fuel will last. Those are separate exercises.

Begin with a census of essential loads: life safety, refrigeration, servers, pumps, lighting in egress routes, and any process that cannot tolerate an unplanned stop. Add the starting current of the largest motor, because a set that carries the running load may still trip on the starting surge. Bank the load in steps and record voltage, frequency, coolant temperature, oil pressure and exhaust temperature at each step.

Run the test long enough for temperatures to stabilise, not just long enough for the meters to look right. A thirty-minute run at half load tells you less than a two-hour run at three-quarter load. Record the result in the same dossier as the engine data, with date, ambient temperature and fuel level at start.

Fuel estimates and the maintenance routine

Fuel consumption is not a single figure. It varies with load, and a set at 25 percent load burns disproportionately more per kWh than one at 75 percent. Build the estimate from the maker’s curve at the loads the building actually expects, then add the starting and cool-down periods. Keep the tank above the level that covers the longest credible outage plus a margin, and treat fuel age as a maintenance item: diesel stored for long periods needs testing and, where required, treatment.

Maintenance follows the engine make and the running hours, not the calendar alone. Oil and filter intervals, coolant condition, belt tension, battery voltage and charger operation, block heater function, and the cleanliness of the radiator and the alternator vents all belong on a checklist. A set that starts reliably in a test but has a weak battery and a blocked radiator will fail in a real outage.

Safety items belong in the same routine. Carbon monoxide from an engine or from a test run in an enclosed space is a hazard that no controller alarm will catch. Return of mains voltage onto a generator still running, or onto a circuit a technician believes is dead, is prevented by the interlocks and by lockout procedure, not by habit. Write both into the dossier and check them at every test.

Keeping the dossier usable

A dossier that lives in a folder nobody opens is decoration. Keep one sheet per engine make with machine code, serial number and documentary version; keep the controller event log exported after each test; keep the transfer settings and the essential load list current; and keep the last load test, fuel estimate and maintenance record together. When a fault appears at two in the morning, the person on the phone needs the engine make, the last alarm and the transfer settings, and nothing else.

The test of a standby system is not whether it starts. It is whether the building can show, from its own records, what the set is, what it did, and what it will carry. That is a documentary exercise as much as a mechanical one, and it starts with the engine make on the first page.