Synchronous Machines & Generators: Locking to the Grid

A magnet that refuses to drift

Picture two dancers in a waltz. One is the rotating magnetic field the stator coils conjure out of three-phase AC — it sweeps around the air gap at a fixed speed set by the grid. The other is the rotor: a powerful magnet, either a wound electromagnet fed DC or a slab of permanent magnets. In an induction motor the rotor always lags a little, panting to catch up — that lag is *slip*. A synchronous machine does something stranger and far more disciplined: its rotor locks onto the rotating field and turns in *perfect lockstep*, with zero slip, forever.

Why does this matter so much? Because *zero slip* means the rotor's speed is welded to the line frequency. On a 50 Hz grid a two-pole machine spins at exactly 3000 rpm; a four-pole one at exactly 1500 rpm — never 2999, never drifting with load. That rock-steady speed is what makes synchronous machines the timekeepers of the entire power system. Every generator on a continent's grid is one of these, all of them turning in invisible mechanical harmony.

Synchronous speed (the one rule that defines the whole machine):

        120 · f
  n  =  ---------     [rpm]
   s       P

  f = line frequency (Hz)        P = number of poles

  50 Hz, 2 poles -> n_s = 3000 rpm    (and exactly 50 rev/s)
  50 Hz, 4 poles -> n_s = 1500 rpm
  60 Hz, 2 poles -> n_s = 3600 rpm
  60 Hz, 4 poles -> n_s = 1800 rpm

  Slip s = (n_s - n) / n_s  =  0   ALWAYS, by definition.
  The rotor turns at n_s exactly, or it does not stay synchronous at all.

Synchronous speed is set entirely by grid frequency and pole count — nothing else. A 2-pole 50 Hz rotor literally makes one full turn every 20 ms.

The load angle: a torsion bar made of magnetism

Here is the most beautiful idea in the whole machine. The rotor and the stator field are coupled not by gears or shafts but by an *invisible magnetic spring*. When the machine spins free with no load, the rotor's north pole sits directly opposite the stator field's south pole — perfectly aligned, like a compass needle resting on north. The angle between them, the load angle δ (delta), is zero.

Now ask the machine to do real work — bolt a pump or a conveyor to the shaft. The rotor doesn't slow down (it *can't*; that would break synchronism). Instead it falls back in angle, dropping a few degrees behind the rotating field while still turning at the very same speed. The magnetic spring stretches. The further the rotor lags, the harder the stretched field pulls it forward — and *that* forward pull is the electromagnetic torque driving your load. Load angle, not speed, is how a synchronous machine signals "I'm working harder."

How torque depends on the load angle delta:

           E·V
  T  =  --------- · sin(delta)      (round-rotor machine)
         omega·X_s

  E   = internal (excitation) EMF, set by rotor field current
  V   = terminal/grid voltage
  X_s = synchronous reactance (the machine's internal impedance)
  delta = load angle between rotor & stator fields

  Torque (or power) vs. load angle:

   T |              .--''''''--.
     |          .-''            ''-.
  T_max...........*..................   <- PULL-OUT at delta = 90 deg
     |       .-'  :                  '-.
     |     .'     :
     |   .'       :  operating point sits
     |  /         :  on the rising part (delta < 90)
     | /          :
     +------------+----------------------> delta
     0           90 deg                 180 deg

Torque rises with sin(δ), peaking at δ = 90°. Normal operation lives on the gentle rising slope, far from the cliff edge.

Pull-out: the edge of the cliff

Stretch any spring far enough and it snaps. The magnetic spring is no different. As you pile on load, δ climbs — 10°, 30°, 60° — and the torque the machine can muster keeps rising. But because torque follows sin(δ), it peaks at δ = 90° and then, perversely, starts *falling*. That peak is the pull-out torque (also called the maximum or breakdown torque). Demand even one newton-metre more than the machine can make at 90°, and the rotor can no longer hold on.

What happens next is dramatic. Past 90°, more lag gives *less* pull, so the rotor falls further behind, which gives even less pull, which makes it fall further still — a runaway. The rotor slips out of step entirely and starts pole-slipping: huge surging currents, violent torque pulsations that can shake a generator off its foundations, and protection relays slamming the machine off the grid within cycles. This is losing synchronism, and on a real power plant it is treated as a serious fault.

Stability margin in practice:

  Rated operating point typically:   delta ~ 20 deg - 30 deg
  Pull-out (theoretical limit):      delta  = 90 deg

  Pull-out torque  T_max = E·V / (omega·X_s)
  Steady-state stability margin = T_max / T_rated  ~  2x to 3x

  Designers leave a big margin so transient swings
  (a sudden fault, a switching surge) don't kick delta
  over 90 deg and trip the unit out of step.

Machines run at 20–30° but are built so pull-out sits 2–3× above the rated load — the headroom that absorbs transient swings.

Run it backwards: now it's a generator

Here is the quiet miracle that powers civilisation: the synchronous machine is *symmetric* in time. As a motor, the grid pushes the rotor and the rotor turns the load, with δ lagging behind the field. Now hook a steam turbine, a hydro penstock, or a diesel engine to the shaft and *push the rotor forward* instead. The load angle flips ahead of the field — and power flows the other way. Mechanical energy in, electrical energy out. The very same iron and copper is now a generator, the alternator at the heart of every power station.

The physics underneath is Faraday's law in its purest dress. As the excited rotor sweeps past the stator windings, its moving magnetic field induces a sinusoidal voltage in each phase — this internally generated EMF is exactly the excitation EMF E we met earlier, and it is a close cousin of the back-EMF that opposes current in any spinning machine. In a generator E is the *cause* of output; in a motor it's the *opposition* to input. Same voltage, read from opposite ends.

One machine, two directions of power — set by the SIGN of delta:

  MOTOR mode                         GENERATOR mode
  ----------                         --------------
  grid drives shaft                  prime mover drives shaft
  rotor LAGS field   (delta < 0)     rotor LEADS field  (delta > 0)
  electrical power IN                mechanical power IN
  mechanical power OUT               electrical power OUT

           grid                              grid
            |  P_elec in                       ^  P_elec out
            v                                  |
      [ SYNCHRONOUS MACHINE ] <----shaft----> [ SYNCHRONOUS MACHINE ]
            |                                  ^
            v  P_mech out                      |  P_mech in
          load                              turbine

  P = (E·V / X_s) · sin(delta)   — same equation, delta just changes sign.

Motor and generator are the same device with the load angle's sign flipped. The power equation never changes; only which way δ leans.

Excitation: the dial that trades reactive power

Synchronous machines have a second, almost magical control knob the induction motor simply doesn't have: the DC field current feeding the rotor. Because that current sets the internal voltage E independently of the shaft load, you can dial E *above* or *below* the grid voltage V — and in doing so you decide whether the machine *makes* reactive power or *soaks it up*. Real power is set by the turbine; reactive power is set by excitation. Two knobs, two outputs, beautifully decoupled.

Over-excited (E > V): the machine pushes out reactive power, like a giant capacitor. It runs at a *leading* power factor and props the grid voltage up — exactly what a sagging grid full of inductive motors needs.
Under-excited (E < V): the machine absorbs reactive power, behaving like an inductor. It runs at a *lagging* power factor and pulls grid voltage down — useful on a lightly-loaded line whose voltage is creeping too high.
Unity excitation (E ≈ V): the machine delivers only real power, drawing or supplying essentially no reactive power — power factor sits at 1.

Synchronising: shaking hands with a continent

You can't just slam a generator's breaker shut onto a live grid. The grid behind that breaker is a million megawatts of spinning steel that will *not* be argued with. Connect out of step and the mismatch hits the machine like a sledgehammer — torque spikes that snap shafts, current surges that vaporise windings. Bringing a new generator online is a careful, four-condition handshake called synchronisation, and operators watch a *synchroscope* whose pointer crawls toward 12 o'clock before they close the breaker at exactly the right instant.

Same frequency. The incoming machine must spin so its electrical frequency matches the grid almost exactly (e.g. both at 50.00 Hz) — adjusted by nudging the turbine governor.
Same voltage magnitude. The generator's terminal voltage must equal the grid's — trimmed with the field excitation.
Same phase sequence. The order in which the three phases peak (A-B-C) must match, or the machine fights the grid catastrophically. This is fixed once, at installation.
Same phase angle. At the instant of closing, the waveforms must line up — the synchroscope pointer at 12 o'clock and drifting slowly. Close, and the magnetic spring gently grabs hold with a near-zero load angle.

Once that breaker closes, the machine is *locked to the grid* and can't drift no matter what — it's now one more dancer in a continental waltz of thousands. Open the throttle and δ creeps forward, exporting power; nothing speeds it up. This same hard lock is what makes synchronous machines so prized, and it sets the stage for the next rung: the permanent-magnet synchronous motor, where we trade the wound field and its DC supply for a rotor full of rare-earth magnets, then steer it with electronics to get synchronous-machine precision in a palm-sized drive.