In short. Breakdown diagnostics is the structured work of turning a generator failure into a confirmed cause before anything is replaced. It starts by classifying the event, for example would not crank, cranked but did not fire, started then stopped, ran with no output or tripped on load, then preserves the alarm and conditions, measures the suspected chain and proves the fault. Resetting the controller and restarting is not diagnosis; it is how the same failure returns a week later.
Five patterns that point in different directions
When a customer in Hyderabad rings to say “the generator is not working”, the first job is to find out which of five quite different things happened. Each pattern narrows the search to a different part of the machine, and none of them justifies the same first repair.
| Pattern | Most likely chain | First measurements |
|---|---|---|
| Will not crank | Battery, terminals, starter, start relay, emergency stop circuit, controller not in auto | Battery voltage at rest and during crank attempt, voltage at starter solenoid, controller mode and active shutdowns |
| Cranks but does not fire | Fuel supply, air in fuel, stop solenoid or ECM fuel enable, cranking speed too low, air intake blocked | Cranking speed, fuel at filter and at pump, stop solenoid pull-in, air filter restriction |
| Starts then stops | Low oil pressure shutdown, coolant temperature, overspeed or underspeed, fuel starvation, sensor or wiring fault, controller timers | The shutdown recorded in the controller log, actual oil pressure with a gauge, fuel level and filter condition, sensor wiring |
| Runs, no or wrong voltage | AVR, excitation loss, rotating diodes, sensing fuse, breaker open, loose connection | Residual voltage at terminals, AVR supply and output, exciter field resistance, breaker status |
| Trips when load is applied | Overload or motor starting, governor response, fuel restriction under demand, turbocharger, protection setting, earth fault on load side | Load current per phase at trip, frequency and voltage dip during load step, controller event with timestamp |
The pattern comes from whoever saw the event, so the questions asked on the phone matter: what did the set do, what did the controller show, what was the load and what has changed recently.
Preserve the evidence before changing anything
The most damaging act at a breakdown is pressing reset. A modern controller holds the alarm, the engine parameters at the moment of shutdown and often a timestamped event log. Resetting clears the screen and may push older events out of memory. Before anyone touches the panel, photograph the display, write down the alarm text, the hour-meter reading and the time, and note the position of every selector and breaker. If it is safe, leave the set in the failed state until the technician arrives.
Separate what was reported from what was measured. “It overheated” may mean a warning lamp, a shutdown, steam from the radiator cap or simply a hot room, and each is a different fault. Reports are a starting point; measurements decide. Our general troubleshooting guide lists what each alarm can and cannot tell you.
Follow the chain, not the alarm
An alarm is the controller’s interpretation of a sensor signal. It can be the correct response to a real machine problem, or the response to a faulty sensor, a chafed wire or a wrong setting. A low oil pressure shutdown with a healthy engine and a failed sender looks identical on the screen to a low oil pressure shutdown with a worn oil pump. The diagnostic step is to measure the real quantity independently, with a mechanical gauge, a thermometer or a meter, and compare it with what the controller believes.
The same discipline applies down the chain. A set that trips on load may have an electrical fault, but the cause is just as often mechanical: a fuel filter that passes enough diesel to idle but not to carry 300 kVA, or a turbocharger that no longer delivers boost. Frequency is a useful signal because engine speed and frequency are locked together on a synchronous alternator; frequency sagging under load says the engine cannot make the power, whereas voltage sagging at steady frequency points at the alternator or AVR. Engine-side causes are pursued under engine service and electrical ones under alternator service.
Do not replace the component named by the alarm
Parts changed on the strength of an alarm name are the most expensive guesswork in the trade. A fail-to-start alarm has sold countless starter motors to sites whose real fault was a corroded battery terminal, and AVRs have been fitted to alternators whose rotating diodes had failed. Our rule is that a part is replaced when a measurement condemns it, or a substitution with a known-good unit proves it, not because the controller mentioned it. Where a part is confirmed faulty, identification and compatibility are handled under parts compatibility verification so the replacement matches the engine or alternator build.
Restoring availability and closing the job
After corrective work, the set must demonstrate the behaviour that failed. If it tripped on load, it is run on load; if it failed to start in auto, a simulated mains failure is performed in auto. The report should state the symptom, the measurements, the confirmed cause, the work done, the verification and anything left open. Temporary measures such as a bypassed sensor or a borrowed battery must be written down with a date for permanent repair; an undocumented bypass is a future breakdown with no clue attached.
When the same fault has occurred more than once, or the cause found does not fully explain the evidence, the job should move to a formal root-cause investigation rather than being closed as a repeat repair.
Related photographs and videos
Common mistakes
- Resetting the controller and restarting before anyone records the alarm, which erases the only evidence of what actually happened.
- Replacing the component named in the alarm text without measuring, which spends money on a part that was never faulty and leaves the real cause in place.
- Declaring the set repaired after a no-load start when the complaint was a trip under load, which guarantees a repeat call at the next outage.
- Treating a sensor reading as a measurement, so a failed sender is diagnosed as an engine fault or a real low oil pressure is dismissed as a sensor error.
- Leaving an undocumented temporary bypass in place, which turns a protection device into a disabled one that nobody remembers.
What to send us
Send the generator make, rating and engine model, a photograph of the controller display with the alarm still showing if possible, a description of exactly what the set did in the order it happened, the load connected at the time, the hour-meter reading and any work done on the set in the last few months. Photographs of the engine bay and the panel interior help us arrive with the right parts. Use the service and AMC enquiry form, or call for an urgent breakdown.
Questions engineers ask
What should I check first when a standby generator does not start during a power cut?
Check that the controller is in auto and shows no active shutdown, then battery voltage, then fuel level. These three account for most failed standby starts. If the set cranks slowly, the battery is the suspect; if it cranks well but will not fire, look at fuel supply and the stop solenoid or fuel enable signal.
Should I reset the alarm and try again?
Not before recording it. Photograph the display, note the alarm text and the hour-meter, then decide. A fail-to-start or low fuel alarm can usually be reset once the cause is addressed. A low oil pressure, high coolant temperature or overspeed shutdown should not be reset and restarted without an engineer confirming the real condition, because repeating the start can finish the engine.
Why does the generator run fine with no load and trip when load is applied?
Because load is what exposes a weak link. Common causes are a partly blocked fuel filter, a turbocharger or air filter problem, a governor that cannot respond, an alternator or AVR fault, an overload from motor starting, or a protection setting tighter than the real load profile. Measuring current, voltage and frequency during the load step separates these.
How can I tell whether an alarm is a real fault or a sensor fault?
Measure the quantity independently. Fit a mechanical oil pressure gauge, read coolant temperature with a contact thermometer or infrared gun, check battery voltage with a meter and compare each with the controller display. If the controller disagrees with the instrument, investigate the sensor and wiring; if they agree, the machine has the problem.
How quickly can a breakdown be attended?
That depends on location, technician availability, site access and whether the likely parts are in stock, so no honest supplier quotes one fixed figure for every case. A good description and photographs sent with the call shorten the visit because the engineer arrives with the right tools and spares.
Technical content reviewed by Srinivas Reddy, Managing Partner, Reddy Generators. Last updated 19 October 2026.