In short. A generator controller alarm tells you what the controller detected, not what failed. Diagnosis means capturing the event log before anyone resets it, confirming the DC supply, checking the sensor or input that raised the alarm against an independent measurement, and tracing any missing action from the controller output through relay, wiring and final device. Controllers themselves fail far less often than the sensors, wiring, batteries and settings around them.
What the controller can and cannot tell you
A modern AMF or synchronising controller monitors engine speed, oil pressure, coolant temperature, battery voltage, alternator and mains electrical values, and digital inputs such as emergency stop and remote start. It compares each against programmed thresholds and delays, drives outputs for fuel, crank, breaker close and alarm, and logs events with a time stamp.
Everything on the screen is derived from an input. A low-oil-pressure shutdown means the controller saw an input consistent with low pressure, which may be genuine, a failed sender, a chafed wire shorted to earth, a wrong calibration curve or a DC supply dip. Treating the message as the diagnosis leads to replacing the controller for a fault in a terminal. The guide to generator controllers covers the architecture these functions sit on.
Preserve the evidence before resetting
The most valuable step happens before any tool is opened. Photograph the display with the alarm showing, then the event log, the readings at the time of the event and the hours counter, and note what the set was doing. Only then reset. Operators under pressure reset repeatedly until the set runs, and on some controllers each reset pushes the oldest events out of the log. Where the controller supports it, download the log and keep it with the service record.
A diagnostic order that finds the fault
The sequence below runs from the most common causes to the least and applies to any make of controller.
| Step | What to check | How | Typical finding |
|---|---|---|---|
| 1 | Identify the system | Controller make, model and firmware; single-line arrangement; any BMS or remote link | Decides which manual and which settings apply |
| 2 | DC supply | Battery voltage at rest, during cranking and running; charger output and ripple | Cranking dip below about 8 V on a 12 V system resets the controller or corrupts readings |
| 3 | The input behind the alarm | Measure the physical quantity independently; measure the sender signal at the controller terminal | Failed sender, earth fault on the sender wire, wrong calibration curve |
| 4 | Wiring and earths | Visual, tug test, voltage drop across connections, engine-to-frame earth strap | Corroded earth strap shifts all analogue readings |
| 5 | Outputs | Command the output in manual or test mode and follow it through relay, wiring and final device | Burnt relay contact, open fuel solenoid, failed breaker coil |
| 6 | Settings and logic | Compare thresholds, delays and timers with engine data and the commissioning record | Threshold changed by someone to clear a nuisance alarm |
| 7 | Controller hardware | Only after steps 2 to 6, by substitution or bench test | Genuine failure is uncommon, usually after lightning or a reversed battery |
DC power is the usual hidden cause
Controllers behave strangely when their supply is poor, which is common on standby sets. A battery that reads 12.6 V at rest can collapse to 7 V during cranking if it is sulphated or its terminals are loose; the controller resets mid-crank and logs a fail to start. A controller that loses its clock, dims during cranking or reboots on load transfer is almost always suffering from supply rather than internal failure. Measure battery voltage with a meter clipped on during an actual start before doing anything else, and see battery and starting-system service for the full starting-chain checks.
Inputs: make the controller prove its reading
Every analogue alarm should be checked against an independent instrument: a mechanical gauge for oil pressure, an infrared thermometer for coolant temperature, alternator frequency for speed (50 Hz on a four-pole machine means exactly 1500 rpm) and a clamp meter for voltage and current. If controller and instrument agree, the condition is real. If they disagree, the fault is in the sender, the wiring or the calibration table.
Resistive senders earth through the engine block, so a corroded earth strap puts resistance in series with every sender and shifts all readings the same way, and a sender of a different make has a different curve that must be matched in the configuration. The controllers and sensors page explains the identification needed to supply the right one.
Outputs, settings and the alarm that misleads
When a commanded action does not happen, measure at each link between the controller output and the mechanical response: output terminal voltage, interposing relay, voltage at the solenoid or coil, and the device itself. A fuel solenoid with a burnt hold coil gives a start that dies after a few seconds; a crank relay with pitted contacts gives an intermittent no-crank. Replacing the controller for either fixes nothing.
Settings are the fault that leaves no hardware evidence. A technician raises the high-temperature threshold to stop a nuisance trip; a year later the engine runs with a half-blocked radiator and nobody is warned until the head gasket fails. Every threshold should match the engine manufacturer’s data and the commissioning record; restore the correct value, then fix the condition that caused the nuisance alarm. The page on AMF control logic describes the timers and thresholds that matter most.
Fail to start
The controller reports fail to start when it does not see its running signal within the crank attempts. That covers a battery that cannot crank, a starter that does not engage, an engine with no fuel, an engine that fires but is not recognised as running because the pickup gap is wrong, and a crank-disconnect speed set too high. Each is a different repair, and the log entry showing whether speed was detected during the attempts separates them.
Replacing a controller correctly
When testing does prove the controller faulty, photograph every configuration page first, including sender curves, timers, thresholds and breaker logic. Confirm the replacement supports the same inputs, outputs and engine interface, especially on electronic engines that communicate over CAN. After fitting, load the configuration, calibrate the senders, simulate each protective shutdown, and perform a full automatic start, transfer and retransfer test. A new controller with default settings on an old engine starts the set and shows a bright display while protecting nothing correctly.
Related photographs and videos
Common mistakes
- Resetting the alarm repeatedly before recording the event log, so the evidence of what happened is overwritten.
- Replacing the controller for a low-oil-pressure shutdown that was a sender wire chafed on the engine, which then trips the new controller as well.
- Raising a temperature or frequency threshold to clear a nuisance alarm, which removes the protection and lets the underlying fault grow into an engine failure.
- Calling a fail-to-start a controller fault without measuring battery voltage during cranking, when the controller was resetting on a collapsed supply.
What to send us
Send a photograph of the controller showing the alarm, photographs or a download of the event log, the controller make and model, the generator nameplate, and a description of what the set was doing when the fault occurred and whether it repeats. Battery voltage during cranking, if you can measure it, rules out the most common cause before a visit. Send everything through the service and AMC enquiry form.
Questions engineers ask
The controller shows low oil pressure but the engine has oil. What is wrong?
Most often the sender or its wiring. Fit a mechanical gauge to confirm actual pressure, then measure the sender resistance at the controller terminal and compare it with the sender curve. A wire shorted to earth, a failed sender or a corroded engine earth strap all read as low pressure. Only if the gauge also reads low do you have an engine problem.
Why does the controller reset or go blank when the engine cranks?
Battery voltage is collapsing below the controller’s minimum during cranking. Causes are a weak battery, loose or corroded terminals, undersized battery cables or a starter drawing excessive current. Measure voltage at the controller terminals during a crank; if it dips below the specified minimum, repair the DC system before suspecting the controller.
Does an obsolete controller have to be replaced?
Not while it works and spares can still be found. It becomes a planning issue when the site needs functions it cannot provide, or when it fails and no identical unit exists. A planned replacement in a service window is far cheaper than an emergency one.
Can I change alarm settings myself to stop a nuisance trip?
You can on most controllers with the right access level, but not until the cause is understood. A nuisance trip is usually a sender fault or a developing engine condition. Correct that, then compare the threshold with the engine manufacturer’s data and the commissioning record, and change it only if it is genuinely wrong.
Technical content reviewed by D. Srinivas Reddy, Managing Partner and Service Engineer, Reddy Generators. Last updated 19 October 2026.