MARATHON SE350 Analogue AVR - Analogue Automatic Voltage Regulator for MARATHON Alternators | OEM Replacement

What the SE350 is. A Analogue Automatic Voltage Regulator built by MARATHON for brushless excitation systems. On paper the specification is a list; in service it is a set of limits. The current the board will hold indefinitely is 3-4 A class; the field resistance window it will drive without exceeding its output voltage is 15 ohm; the tightest voltage band it can maintain is +/-1.0 per cent; and the ambient range in which all of that remains true is -40 to +60 degC.

Excitation architecture. PMG or shunt fed depending on variant. That single sentence decides most of the regulator's behaviour in a fault: whether the field survives a close-up short circuit, whether the machine builds up from residual voltage on its own, and how much the terminal voltage dips when a large motor starts on the same bus. Counterfeit regulators are a real problem on the widely copied boards. Buy from a source that will state where the board was made.

Sensing. Single-phase sensing with voltage taps. The sensing circuit is the regulator's only view of the outside world, so its wiring matters as much as its specification. A voltage error introduced by a loose terminal or a long run of untwisted cable is indistinguishable, from the regulator's point of view, from a genuine voltage error on the alternator. A regulator rated for a minimum field resistance of a few ohms will not drive a high resistance field to full current, however good its specification looks on paper.

Power stage and field drive. Power input From the main stator or PMG, 50 / 60 Hz, output 63 Vdc class. Field current rating 3-4 A class, forcing capability Short-term forcing current available for load steps, field resistance window 15 ohm. These are not independent numbers: the output voltage ceiling and the field resistance together set the maximum current the board can actually deliver, and the continuous rating is only reachable at the low end of the resistance window. Take a photograph of the terminal block before removing anything. Wiring drawings for older machines are frequently wrong, and the photograph is the only reliable record.

Accuracy and response. +/-1.0 per cent steady-state accuracy, Standard analogue response, thermal drift Low thermal drift. In practice the accuracy figure is the least interesting of the three. What an operator notices is recovery after a load step, which is governed by response and by available forcing current, and what a maintenance engineer notices after two years is drift, which is governed by the reference circuit and the ambient temperature inside the terminal box. A regulator is a closed loop with a slow plant attached, and the plant is a large inductance. That is why the stability adjustment exists and why it is set after, not before, the voltage is correct.

Frequency behaviour. Rated for 50 / 60 Hz, with Volts-per-hertz roll-off. The frequency-sensitive function exists to protect the prime mover, not the alternator: as engine speed falls the regulator reduces excitation so the alternator stops demanding torque the engine cannot supply. Without it, an overloaded engine is dragged down until it stalls, and the alternator is blamed for a fault that belongs to the fuel or governor system. Vibration kills more regulators than heat does. If a board has failed twice in the same machine, look at the mounting before looking at the supplier.

Paralleling. Reactive droop via CT. On a single set this circuit does nothing. On two or more sets it is the only thing that stops the machines from exchanging reactive power in a slow oscillation, and it is the first place to look when a parallel installation shows a circulating current that nobody can explain. The trim potentiometer is not a calibration device. If rated voltage is not reachable near the middle of its range, the sensing tap is wrong.

Protection built into the regulator. Over-excitation shutdown. These functions protect the alternator and the regulator, not the installation. Downstream protection, loss of mains and reverse power all belong to the controller, and a site that relies on the AVR for them is a site that will eventually trip in a way nobody can explain. Any AVR that has been in service for more than ten years is running on borrowed time, and the failure always happens at the least convenient moment.

What it fits. Marathon small and mid-size alternators. Lifecycle status: In production. Both halves of that sentence matter at order time. A board that fits perfectly but has been discontinued is a different commercial decision from one that fits and is in current production, and the price difference rarely reflects the difference in risk. Never flash a field through the regulator. Disconnect the board, flash the field directly, then reconnect. Flashing through the AVR is how a working board becomes a failed board.

Field notes on this model. This is the analogue regulator used on Marathon's smaller alternator frames and as a service replacement on older sets. It is a straightforward board with three adjustments, and the most common field failure is a dried-out electrolytic rather than a semiconductor, which is why a fifteen-year-old unit that has lost regulation often comes back to life with a re-cap.

What the customer is actually buying. When a regulator is quoted as a replacement, the buyer is not purchasing a circuit board. They are purchasing the certainty that the set will hold voltage on the next load step. On the SE350 that certainty comes from the combination of 3-4 A class continuous field current and the Standard analogue response response, and it disappears the moment the field resistance drifts outside the window the board was set up for.

Field cable sizing. The field circuit carries DC at low voltage and high current, so volt drop matters more than ampacity. Size the cable so that the drop at 3-4 A class stays under about one volt, and torque the terminals to the manufacturer figure. A field circuit that loses half a volt in the cable is a regulator that cannot reach full forcing when the load steps.

Commissioning sequence. Start with the machine at rated speed, no load, and the field disconnected. Confirm the sensing voltage at the AVR terminals is inside the window the board expects, then connect the field with the voltage trim at minimum. Bring the trim up slowly and watch the field current, not the terminal voltage, because the terminal voltage will look correct long before the field current has stabilised on a cold winding.

Measure, do not assume. Nameplate field resistance figures are nominal at 20 degC and are frequently wrong after a rewind. Take the measurement with a four-wire meter if one is available, because a two-wire reading on a low resistance winding includes the lead resistance and can be off by twenty per cent.

Symptom first, board second. A set that will not build up voltage has one of four causes: no residual magnetism, a broken sensing circuit, a failed regulator, or a field circuit open. Test them in that order, because the first two are free to check and the third is not. Flashing the field with a battery through a resistor will separate a residual problem from a regulator problem in under a minute.

Counterfeits. Where a board is widely copied, buy from a source that will state the country of manufacture and stand behind it. A counterfeit regulator usually works for a month and then fails on the first cold morning, and by then the machine has a reputation problem rather than a parts problem.

Parallel operation is a different animal. A single set on an island load only has to hold voltage. Two sets in parallel also have to agree about reactive load, and that agreement is made or broken by the droop circuit, not by the voltage accuracy. If a site is going to parallel, the regulator choice has to be made with the droop method in mind from the beginning, because retrofitting load sharing onto a board that was never specified for it costs more than buying the right one.

Field cable sizing. The field circuit carries DC at low voltage and high current, so volt drop matters more than ampacity. Size the cable so that the drop at 3-4 A class stays under about one volt, and torque the terminals to the manufacturer figure. A field circuit that loses half a volt in the cable is a regulator that cannot reach full forcing when the load steps.