An electric generator is a rotating electromagnetic machine that converts mechanical shaft power into electrical power by inducing voltage in a set of conductors moving relative to a magnetic field. The core physics is Faraday's law — a changing flux linkage through a coil produces an EMF proportional to the rate of change. Construction centres on a laminated stator core carrying the armature winding, a rotor carrying the field winding or permanent magnets, and a tightly controlled air gap. Real machines like the Siemens SGen-100A-2P deliver up to 370 MVA from this same basic layout.
Electric Generator Construction Interactive Calculator
Vary the generator air-gap range, bore size, tolerance rate, and rotor speed to see clearance, tolerance, and 2-pole frequency update on the cross-section.
Equation Used
This calculator uses the article air-gap construction values. The selected 50-80 mm gap range is reduced to a nominal midpoint, the bore concentricity allowance is calculated from the tolerance rate in mm per metre of bore diameter, and 2-pole synchronous frequency is calculated from rotor speed.
- 2-pole synchronous generator cross-section.
- Nominal radial air gap is the midpoint of the selected low and high gap values.
- Bore concentricity tolerance is applied as c times bore diameter.
- Magnetic saturation, winding factor, and detailed EMF sizing are not included.
How Does an Electric Generator Actually Work?
A generator works because a coil sees a changing magnetic flux. Spin a magnetised rotor inside a stator wound with copper conductors and you induce an EMF in those conductors. The waveform you get out depends on how the field is shaped, how the slots are skewed, and how the windings are distributed around the bore. In a 3-phase synchronous machine the stator carries three winding sets spaced 120 electrical degrees apart, and the rotor carries a DC-excited field winding (or permanent magnets) that produces the rotating flux as the shaft turns.
The construction is built around one number you cannot fudge — the air gap. On a 50 MW turbo-generator the radial air gap sits around 50-80 mm. On a small 10 kW PMG it might be 1.5 mm. Push the gap too wide and you lose flux density in the gap, the EMF drops, and you have to dump more current into the field winding to compensate, which heats the rotor. Run it too tight and rotor thermal growth, shaft sag, or bearing wear lets the rotor rub the stator bore — that's a stator core grind, and on a large machine it ends the unit. Tolerances on stator bore concentricity typically run ±0.05 mm per metre of bore diameter on a quality build.
"The air gap is the one number you cannot fudge in a generator build. Hold concentricity to ±0.05 mm per metre of bore and the bearings stay happy; let it drift and unbalanced magnetic pull starts loading them at twice line frequency until something gives." — Robbie Dickson, FIRGELLI Automations founder and former Rolls-Royce, BMW, and Ford engineer
Failure modes track straight back to construction. Insulation breakdown on the stator winding from partial discharge in the slot, copper bar vibration loosening wedges, rotor pole-face arcing from contaminated slip rings, and core lamination shorts from a damaged interlaminar varnish — every one of these is a construction-quality issue, not an operating one. That's why good generator construction obsesses over lamination stacking pressure (typically 1.5-2.0 MPa), winding insulation class (F or H on most modern machines), and end-winding bracing to handle short-circuit forces that can hit 100× normal current.
Key Components
- Stator Core: Stack of thin silicon-steel laminations, typically 0.35-0.50 mm thick, coated with interlaminar varnish to suppress eddy currents. The slots punched in the bore hold the armature winding. Stacking pressure must hit 1.5-2.0 MPa to prevent vibration-induced lamination chatter that erodes the varnish and shorts the core.
- Stator Winding (Armature): Distributed 3-phase copper winding sitting in the stator slots, typically Class F or H insulation rated for 155 °C or 180 °C hot-spot (per IEC 60034-1 and NEMA MG 1). On large machines the bars are Roebel-transposed to minimise eddy losses. End-winding bracing must withstand fault forces of 100× rated current without permanent deformation (per IEEE 115 and IEC 60034-3).
- Rotor (Field System): Either a salient-pole rotor for low-speed hydro machines (typically 100-750 RPM) or a cylindrical rotor for 2-pole and 4-pole turbo-generators (1500, 1800, 3000, or 3600 RPM). Carries the DC field winding excited through slip rings, or permanent magnets in a PMG.
- Air Gap: The radial clearance between rotor outer surface and stator bore. Sets the magnetising current and the synchronous reactance. Concentricity tolerance is ±0.05 mm per metre of bore — beyond that you get unbalanced magnetic pull which loads the bearings and causes 2× line-frequency vibration (assessed per ISO 20816-2 for large turbine generator sets).
- Slip Rings and Brushes: On a wound-field machine, two slip rings on the rotor shaft carry the DC field current in through carbon brushes. Brush pressure runs 18-25 kPa for stable contact. Lose pressure and you arc the rings, contaminate the surface, and lose excitation control.
- Exciter: Provides the DC field current — either a brushless exciter (small AC generator with rotating rectifier on the same shaft) or a static excitation system fed from the terminals through a transformer and thyristor bridge. Sized to deliver typically 0.5-1% of rated machine power at full field.
- Bearings and Shaft: On large machines, hydrodynamic sleeve bearings with forced lube oil; on smaller units, sealed deep-groove ball bearings. Shaft must hold runout under 0.025 mm TIR at the rotor body to keep the air gap concentric.
- Cooling System: Air-cooled for machines below ~50 MW, hydrogen-cooled for 50-500 MW (hydrogen has 7× the heat-transfer coefficient of air at 4 bar), and water-cooled stator bars on the largest units. Loss of cooling drops insulation life — every 10 °C above rated hot-spot halves winding life (the Arrhenius-type 10 K rule used in IEEE and IEC insulation system standards).
Generator design starts with geometry, not megawatts. The air gap, the bore concentricity, and the shaft runout decide whether the machine survives — the rating only tells you how hard it will fail.
Where Are Electric Generators Used in the Real World?
Generator construction varies wildly with the application. A 1500 MVA turbo-generator at a nuclear plant and a 5 kW diesel genset on a fishing boat share the same physics but almost nothing in their build. The choice of salient-pole vs cylindrical rotor, brushless vs slip-ring excitation, and air vs hydrogen cooling all come down to speed, rating, and duty cycle. You would be amazed how much of the design tree falls out of just two numbers — rated power and prime-mover speed.
- Utility Power Generation: Siemens SGen5-2000H hydrogen-cooled 2-pole turbo-generator at 3000 RPM driving up to 370 MVA, paired with a steam turbine in combined-cycle plants like Irsching 4 in Germany.
- Hydroelectric: GE salient-pole synchronous generators at the Itaipu Dam, 700 MW per unit at 90.9 RPM with a 16 m diameter rotor — the salient pole construction handles the low speed and the huge inertia.
- Wind Power: Vestas V164 8 MW direct-drive permanent magnet generator with no gearbox — a 5+ m diameter PMG running at 12 RPM rotor speed.
- Backup and Standby: Cummins QSK60 diesel-driven 2 MW genset with a brushless 4-pole synchronous alternator at 1800 RPM for hospital and data centre standby.
- Marine: Caterpillar 3516 marine genset with a Kato Engineering 6-pole alternator at 1200 RPM running on heavy fuel oil, sized 1825 ekW for tug and OSV propulsion-assist duty.
- Aerospace: Honeywell 90 kVA integrated drive generator on the Boeing 737 — a constant-speed-drive feeding a 3-phase 400 Hz brushless alternator at 24,000 RPM.
- Portable Power: Honda EU7000iS inverter generator with a multi-pole permanent magnet alternator running at variable speed, rectified to DC then inverted to clean 60 Hz output.
What Formula Governs Generator EMF and Sizing?
The EMF equation tells you the RMS voltage induced per phase as a function of flux per pole, frequency, turns per phase, and the winding factor. It's the equation you use to size the stator winding once you've fixed the magnetic loading. At the low end of a typical turbo-generator design you might run 1.4 T peak air-gap flux density and a winding factor near 0.92 — that gets you a conservative, easy-to-cool machine. At the high end you push 1.8 T and a winding factor of 0.96, which packs more voltage per turn but drives the iron close to saturation and the losses up sharply. Most utility machines sit in the sweet spot at 1.55-1.65 T peak.
Variables
| Symbol | Meaning | Unit (SI) | Unit (Imperial) |
|---|---|---|---|
| Eph | RMS EMF induced per phase | V | V |
| f | Electrical frequency (pole pairs × shaft rev/s) | Hz | Hz |
| Φ | Magnetic flux per pole | Wb | Wb (or Mx × 10-8) |
| Nph | Series turns per phase | turns | turns |
| kw | Winding factor (distribution × pitch) | dimensionless | dimensionless |
Worked Example: Electric Generator Construction in a 2 MW biogas CHP generator
Sizing the stator winding of a 2 MW 4-pole synchronous generator coupled to a Jenbacher J420 biogas engine running at 1500 RPM on a European 50 Hz grid. The machine has a stator bore of 700 mm, axial core length of 600 mm, and you need a phase EMF of 6350 V (which gives 11 kV line-to-line in star). You're choosing series turns per phase given a target air-gap flux density.
Given
- f = 50 Hz
- Eph = 6350 V
- kw = 0.94 dimensionless
- Bpeak = 1.6 T (nominal)
- Pole pitch area = 0.165 m² (per pole)
Solution
Step 1 — compute flux per pole at nominal 1.6 T peak air-gap flux density. The average flux density over a sinusoidal pole is (2/π) × Bpeak, and Φ = Bavg × Apole:
Step 2 — solve the EMF equation for series turns per phase at the nominal flux:
Step 3 — at the low end of typical practice, Bpeak = 1.4 T (a conservative cool-running design like a continuous-duty marine alternator):
That extra 26 turns means more copper in the slots, lower flux density, cooler iron, and easier voltage regulation — but you pay in copper cost and slightly higher I²R loss for the same output current. Step 4 — at the high end, Bpeak = 1.8 T (aggressive utility-grade turbo-generator territory):
Fewer turns, more compact winding, but the iron is close to the knee of the B-H curve. Magnetising current jumps, no-load core losses can climb 30-40%, and you'll hear the laminations buzzing at 100 Hz under no-load magnetisation. The 1.6 T nominal sits right where most 2 MW gensets actually get built — Stamford and Leroy-Somer machines in this class typically land within 5% of this turn count.
Result
The nominal design needs 181 series turns per phase to deliver 6350 V phase EMF at 50 Hz. That number sets your slot count, conductor cross-section, and ultimately the slot fill factor — typically 0.45-0.55 for a random-wound machine in this size class. The low-end 1.4 T design needs 207 turns and runs cooler with more regulation margin; the high-end 1.8 T design squeezes down to 161 turns but flirts with iron saturation. If you build the machine and measure phase EMF 8% below predicted at rated field current, three failure modes lead the suspect list: (1) a stator core stacking pressure below 1.5 MPa allowing axial gaps that kill effective flux area, (2) a winding pitch error of one slot which drops kw from 0.94 to around 0.88, or (3) excessive air gap from a machined-undersize rotor body — a 0.5 mm radial overcut on a 2 mm nominal gap drops air-gap flux density by roughly 20%.
How Does Generator Construction Compare Across Synchronous, Induction, and PMG Types?
Generator construction is not one design — it's a family. The choice between synchronous, induction, and permanent magnet construction comes down to speed, control needs, fault behaviour, and how clean the prime-mover speed is. Here's how the three main constructions stack up on the dimensions that actually matter when you're specifying a machine.
| Property | Synchronous Generator (wound field) | Induction Generator | Permanent Magnet Generator (PMG) |
|---|---|---|---|
| Typical power range | 1 kW – 1500 MVA | 10 kW – 5 MW | 100 W – 10 MW |
| Voltage regulation method | Active via field current (AVR) | Reactive power from grid or capacitor bank | Fixed flux — needs power electronics for regulation |
| Speed-frequency coupling | Locked (synchronous) | Slips 1-3% under load | Variable speed, requires rectifier/inverter |
| Construction complexity | High — slip rings, exciter, AVR | Low — squirrel cage, no brushes | Medium — magnets are simple, but assembly is hazardous |
| Efficiency at rated load | 96-99% on large machines | 92-96% | 94-97% direct-drive |
| Fault current contribution | High and sustained (5-7× rated) | Decays in 3-5 cycles | Limited by inverter (typically 1.1-1.5× rated) |
| Capital cost per kW (2 MW class) | ~$80-120/kW | ~$60-90/kW | ~$150-250/kW (NdFeB magnet cost) |
| Best application fit | Grid-tied utility, large gensets | Small hydro, retrofits, simple gensets | Wind direct-drive, variable-speed inverter gensets |
Frequently Asked Questions About Electric Generator Construction
Almost always a winding error — most commonly a coil pitch shortened by one slot, which moves kw in the wrong direction, or extra turns put in by mistake during the rewind. Count the conductors per slot against the original spec sheet and verify the coil span with a slot-to-slot test before you blame the iron.
Less common but worth checking: if the stator was re-stacked after a rewind and the core stacking pressure ended up higher than the original 1.5-2.0 MPa target, the effective iron area increases slightly and flux per pole goes up. You'll see this as both higher EMF and slightly lower no-load magnetising current.
Pick brushless when you want low maintenance and you can live with slower field response — typically 0.5-1.0 second to ceiling voltage. The rotating rectifier handles the field current with no slip rings and no carbon dust, which matters in dusty cement plant or grain mill installations.
Pick static excitation when you need fast field forcing for grid stability — response times under 50 ms. The trade-off is slip rings and brushes that need inspection at 4000-8000 hour intervals, plus a field flashing circuit because the static system can't bootstrap from residual magnetism alone.
The most likely cause is shorted turns in the field winding itself. A few shorted turns on the rotor reduce the effective ampere-turns per pole, so the AVR ramps up field current to compensate and still hits the voltage target — but you'll see it on the field ammeter as a step-up from baseline. Run a recurrent surge oscilloscope (RSO) test on the rotor to confirm.
Second possibility: degraded interlaminar insulation in the stator core causing localised eddy current heating and lowered effective permeability. You'll see this as warm spots on the back-of-core temperature scan during heat-run testing.
Start with a radial air gap of roughly 0.1% of the stator bore diameter as a baseline — for a 300 mm bore, that's 0.3 mm, which is too tight for a real build. Then add manufacturing tolerance stack-up (rotor TIR, bearing clearance, thermal growth) and you typically land at 1.0-1.5 mm for a small PMG.
The trade-off: smaller gap means higher air-gap flux density for the same magnet thickness, but cogging torque and tolerance sensitivity climb sharply below 1 mm. If you go below that you need ground rotor and stator surfaces, not just machined.
The textbook calculation uses synchronous reactance Xs, but during the first few cycles after a load step the machine actually responds with subtransient reactance Xd'' (very low, big initial current and dip), then transient reactance Xd' (moderate), and only after several seconds settles to Xs. If your AVR is slow, the dip you measure on a scope is dominated by Xd' not Xs.
If the dip recovery is also sluggish, check AVR field-forcing ceiling — many off-the-shelf AVRs only force to 1.6× rated field voltage, which limits how fast they can pull voltage back up after a heavy motor start.
Spec Class F (155 °C) for continuous duty and run it at Class B temperature rise (80 K). That gives you 25 K of thermal margin, which translates to roughly 4× the insulation life of running Class F at full Class F rise. Most reputable utility-class machines are built this way regardless of nameplate.
Class H (180 °C) makes sense only if you genuinely need the thermal headroom — high ambient (above 50 °C), altitude derating, or duty-cycle peaks above continuous rating. Otherwise you pay for insulation system you'll never use, and Class H varnishes can be more brittle in service.
Twice line frequency (2f) vibration on a synchronous generator points straight at unbalanced magnetic pull from a non-concentric air gap. After an overhaul the most likely cause is the stator being re-set on its sole plates with a slightly off-centre bore, or a bearing pedestal shimmed unevenly. Spec is typically ±0.05 mm per metre of bore — measure with a feeler gauge at four positions around the rotor.
If the air gap is concentric, look at the rotor itself — a shorted field turn on one pole creates an asymmetric flux distribution that also shows up as 2f vibration, and unlike mechanical asymmetry this one comes and goes with field current.
References & Further Reading
- Wikipedia contributors. Electric generator. Wikipedia
Building or designing a mechanism like this?
Explore the precision-engineered motion control hardware used by mechanical engineers, makers, and product designers.
