Skip to content
LumenaryAcademy
Power 75 min

Transformers and electric machines

How magnetic fields change voltages and turn shafts, from the distribution transformer down the street to the motors in electric cars and drones.

Code tested · 5AI-draftedAwaiting review by an electrical engineer

Your guides: Chloe · The Physicist, Margaret · The Bench Engineer, Gideon · The Design Reviewer

MotivationH

Machines that run the world

In 2011 the International Energy Agency estimated that electric motors and the systems they drive (pumps, fans, compressors, conveyors) use more than 40% of the world's electricity. Add the generators that make nearly all of that electricity, and the transformers it passes through, typically several times, between a power station and a plug, and you have the machines this lesson is about. They all work by one principle, discovered by Michael Faraday in 1831: a changing magnetic flux through a coil induces a voltage in it, and its mirror image, a current in a magnetic field feels a force.

A transformer uses the first half alone, with nothing moving. A motor uses both: currents in fields make torque, and the motion induces a voltage that pushes back. Learn the equivalent circuits of three or four machines and you can predict how a distribution transformer heats up through the day, why a motor draws six times its rated current when it starts, and why the motors in electric cars are shaped by the price of rare-earth magnets.

Note

Safety. Transformers and motors connected to the mains or to a vehicle's traction battery carry lethal voltages, and large machines store enough energy in their rotation and their magnetic fields to injure. This lesson is calculation and simulation. If you want to see a motor and its back-EMF for yourself, a small hobby DC motor on a few AA cells or a current-limited bench supply at a few volts is safe; never work on mains-powered motors, transformers or EV drive systems.

DefinitionsH

The transformer

Wind two coils, of N1N_1 and N2N_2 turns, on one closed core of laminated steel. An AC voltage on the first drives a changing flux Φ(t)\Phi(t) round the core, and by Faraday's law the same flux links both coils: v1=N1 dΦ/dtv_1 = N_1\,d\Phi/dt and v2=N2 dΦ/dtv_2 = N_2\,d\Phi/dt.

Definition— Ideal transformer

An ideal transformer has no winding resistance, no leakage flux, no core loss, and a core so permeable that it needs no magnetising current. Its voltages and currents obey v1v2=N1N2,N1i1=N2i2,\frac{v_1}{v_2} = \frac{N_1}{N_2}, \qquad N_1 i_1 = N_2 i_2, the second because the core's net magnetomotive force must be zero. Power in equals power out, v1i1=v2i2v_1 i_1 = v_2 i_2, and an impedance ZZ on the secondary looks like Z′=(N1/N2)2ZZ' = (N_1/N_2)^2 Z from the primary: impedance is reflected by the square of the turns ratio.

For a sinusoidal flux of peak value Φmax⁡\Phi_{\max}, the RMS voltage of a coil is V=2π2fNΦmax⁡≈4.44 fNΦmax⁡V = \tfrac{2\pi}{\sqrt2} f N \Phi_{\max} \approx 4.44\,fN\Phi_{\max}, the EMF equation. Turn it round: for a given voltage and frequency, the peak flux, and so the peak flux density B=Φ/AB = \Phi/A in the core, is fixed. Electrical steel saturates at about 1.5 to 1.8 T, so a designer picks the core area and number of turns to keep BB below that. It also explains why a transformer only works on AC (a constant voltage would make the flux grow without limit until the core saturated and the winding resistance alone limited the current), and why a 60 Hz transformer run at 50 Hz overheats: its flux, and its magnetising current, rise.

A real transformer departs from the ideal in four ways, each modelled by one element in its equivalent circuit:

  • Winding resistance R1R_1, R2R_2: copper loss I2RI^2R, which grows with the square of the load.
  • Leakage reactance X1X_1, X2X_2: flux that links one winding but not the other, which acts as a series inductance and makes the output voltage sag under load.
  • Magnetising reactance XmX_m: the finite permeability of the core, which needs a small magnetising current (a few per cent of rated current, or less) even at no load.
  • Core-loss resistance RcR_c: hysteresis loss (the energy spent re-orienting the steel's magnetic domains every cycle) and eddy-current loss (currents induced in the core itself). Cores are built of thin insulated laminations of grain-oriented silicon steel, or amorphous metal, to keep both small. Core loss depends on voltage and frequency, not on load, so it is there all day and all night.
R1jX1R2′jX2′RcjXmN1 : N2idealloadV1
The equivalent circuit of a real transformer, with the secondary's resistance and leakage reactance referred to the primary side. The series elements carry the load current and cause copper loss and voltage drop; the shunt branch carries the magnetising current and represents core loss. On the right, an ideal transformer restores the actual secondary voltage.

Transformers are among the most efficient machines people build. Their efficiency at a load fraction xx (of rated kVA) and power factor pf\mathrm{pf} is η=xS pfxS pf+Pcore+x2Pcu,\eta = \frac{x S\,\mathrm{pf}}{x S\,\mathrm{pf} + P_{\text{core}} + x^2 P_{\text{cu}}}, where PcuP_{\text{cu}} is the copper loss at full load. Differentiating shows that efficiency peaks where copper loss equals core loss, at x=Pcore/Pcux = \sqrt{P_{\text{core}}/P_{\text{cu}}}. Distribution transformers spend most of their lives lightly loaded, so they are designed with low core loss and reach their best efficiency well below full load.

python
import numpy as np

# An illustrative 400 kVA, 11 kV / 400 V distribution transformer.
S_rated = 400e3            # VA
P_core = 0.6e3             # W: core (no-load) loss, present whenever the transformer is energised
P_cu_full = 4.5e3          # W: copper (load) loss at rated current; it scales with current squared
z_pu = 0.045               # its short-circuit impedance: 4.5% on its own rating

print("  load    pf   output kW   core W   copper W   efficiency")
for x in [0.1, 0.25, 0.365, 0.5, 0.75, 1.0, 1.2]:
    for pf in [1.0, 0.8]:
        P_out = x * S_rated * pf
        P_cu = x**2 * P_cu_full
        eta = P_out / (P_out + P_core + P_cu)
        print(f"{x:6.1%}  {pf:4.1f}  {P_out / 1e3:9.1f}  {P_core:7.0f}  {P_cu:9.0f}   {eta:9.2%}")
x_best = np.sqrt(P_core / P_cu_full)
print(f"maximum efficiency at {x_best:.1%} of rated load, where copper loss equals core loss")

# Voltage regulation: how far the secondary voltage sags from no load to full load.
r = P_cu_full / S_rated                  # winding resistance in pu = full-load copper loss / rating
x_l = np.sqrt(z_pu**2 - r**2)            # leakage reactance in pu
print(f"series impedance {r:.4f} + j{x_l:.4f} pu")
for pf, name in [(1.0, "pf 1.0"), (0.8, "pf 0.8 lagging")]:
    I = pf - 1j * np.sqrt(1 - pf**2)     # rated current (1 pu) with 1 pu voltage at the load
    V_source = 1 + I * (r + 1j * x_l)    # the voltage needed behind the series impedance
    print(f"{name:15}: regulation {abs(V_source) - 1:.2%}")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

The transformer is above 98% efficient from a tenth of its load to 20% overload, and peaks at 99.19% at 36.5% load. The voltage regulation, the fall in secondary voltage from no load to full load, is 1.22% at unity power factor but 3.55% at 0.8 lagging, because the lagging current's drop across the leakage reactance lines up with the voltage. That is one more reason utilities dislike low power factors, and why distribution transformers have tap changers, which adjust the turns ratio by a few per cent in steps.

ComputationC

The DC motor: torque and back-EMF

A DC motor has a stationary field, from permanent magnets or a field winding, and a rotating armature winding fed through a commutator, a segmented ring with sliding brushes that reverses each coil's current as it turns so that the torque always pushes the same way. Two equations describe it, and they share one constant.

Definition— DC motor equations

With armature current ii, angular speed ω\omega and motor constant kk: T=k i(torque),e=k ω(back-EMF).T = k\,i \quad\text{(torque)}, \qquad e = k\,\omega \quad\text{(back-EMF)}. The armature circuit and the shaft obey v=R i+Ldidt+kω,Jdωdt=k i−b ω−Tload,v = R\,i + L\frac{di}{dt} + k\omega, \qquad J\frac{d\omega}{dt} = k\,i - b\,\omega - T_{\text{load}}, where RR and LL are the armature's resistance and inductance, JJ the inertia of rotor and load, and bb a friction coefficient.

In steady state, v=Ri+kωv = Ri + k\omega and T=kiT = ki give a straight-line torque–speed curve: ω=(v−RT/k)/k\omega = (v - RT/k)/k. At no load the motor spins at about v/kv/k, where its back-EMF nearly equals the supply. Stalled (ω=0\omega = 0), nothing opposes the supply but RR, so the current is v/Rv/R, often ten or more times the rated current. Starting a motor is therefore a brief stall: the back-EMF that limits the current grows only as the motor gathers speed.

The simulation integrates both equations for a small 24 V motor of the size that drives a robot's wheel, first switched straight onto the supply, then fed by a drive that limits the current to 15 A by lowering the voltage while the motor is slow.

python
import numpy as np

# A small permanent-magnet DC motor, the size that drives a robot's wheel (illustrative values).
V = 24.0        # supply voltage, V
R = 0.6         # armature resistance, ohm
L = 0.8e-3      # armature inductance, H
k = 0.06        # torque constant in N m/A, equal to the back-EMF constant in V s/rad
J = 5e-4        # inertia of rotor and load, kg m^2
b = 2e-5        # viscous friction, N m s
T_load = 0.3    # load torque, N m

def simulate(i_limit=None, t_end=0.5, dt=1e-5):
    """Integrate  L di/dt = v - R i - k w  and  J dw/dt = k i - b w - T_load.
    With i_limit, a drive lowers the applied voltage so that the current stays below i_limit."""
    i = w = 0.0
    rows, i_peak, t95 = [], 0.0, None
    w_final = (k * V - R * T_load) / (k**2 + R * b)        # where both derivatives are zero
    every = int(round(0.05 / dt))
    for n in range(int(round(t_end / dt)) + 1):
        if n % every == 0:
            rows.append((n * dt, i, w * 60 / (2 * np.pi)))
        v = V if i_limit is None else min(V, k * w + R * i_limit)
        i += dt * (v - R * i - k * w) / L
        net = k * i - b * w - T_load
        if w <= 0 and net < 0:                             # the load cannot drive the motor backwards
            net = 0.0
        w += dt * net / J
        i_peak = max(i_peak, i)
        if t95 is None and w >= 0.95 * w_final:
            t95 = (n + 1) * dt
    return rows, i_peak, t95, w_final

for label, lim in [("switched straight onto 24 V", None), ("drive limiting the current to 15 A", 15.0)]:
    rows, i_peak, t95, w_final = simulate(lim)
    print(f"{label}: peak current {i_peak:.1f} A, 95% of final speed after {t95 * 1000:.0f} ms")
    print("  t (ms)   current (A)   speed (rpm)")
    for t, i, rpm in rows:
        print(f"  {t * 1000:6.0f}   {i:11.2f}   {rpm:11.0f}")
print(f"final speed {w_final * 60 / (2 * np.pi):.0f} rpm, final current {(T_load + b * w_final) / k:.2f} A;"
      f" stall current V/R = {V / R:.0f} A")
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

Switched straight on, the current leaps to 38.2 A within about 6 ms (nearly the stall current of 40 A, held back only by the inductance), then falls as the back-EMF builds, settling at 5.12 A and 3331 rpm. With the current limited to 15 A the motor accelerates at constant torque, a straight ramp in speed, and reaches 95% of its final speed in 353 ms instead of 246 ms. Every drive makes this trade: a gentler start, kinder to the windings, the gears and the supply, at the cost of a slower one. The drive does it with the pulse-width modulation of Power electronics, and the speed loop around it is the controller of Feedback control.

Worked examplesH

The induction motor and slip

The rotating field of AC power and three-phase systems turns at the synchronous speed ns=120fp rpmn_s = \frac{120 f}{p}\ \text{rpm} for a stator wound with pp poles: 3000 rpm for two poles at 50 Hz, 1500 for four, 1200 for six at 60 Hz. Inside it sits a rotor of laminated steel with aluminium or copper bars along it, shorted together by rings at each end, the squirrel cage. Nothing is connected to the rotor. The field sweeping past its bars induces currents in them, and those currents, in the field, feel a force that drags the rotor along.

Definition— Slip

The slip of an induction motor turning at nn rpm is s=ns−nns.s = \frac{n_s - n}{n_s}. At standstill s=1s = 1; at synchronous speed s=0s = 0. The rotor's currents alternate at the slip frequency sfsf.

If the rotor turned at exactly synchronous speed, its bars would move with the field, see no change of flux, carry no current and make no torque. So an induction motor always runs slightly slower than synchronous, by just enough slip for the rotor currents to produce the torque the load demands: typically 1 to 5% at full load. More load, more slip.

The per-phase equivalent circuit looks like a transformer's, because it is one, with a secondary that rotates. The one change is that the rotor resistance appears as R2′/sR_2'/s, and the power that resistance absorbs, the air-gap power, is what crosses to the rotor. A fraction ss of it is lost as heat in the rotor bars; the remaining 1−s1 - s becomes mechanical power. So the torque is T=3 I2′2 R2′/sωs,T = \frac{3\,I_2'^2\,R_2'/s}{\omega_s}, with ωs\omega_s the synchronous speed in rad/s. Replacing the supply and stator by their Thévenin equivalent Vth\mathbf V_{\text{th}}, Zth=Rth+jXthZ_{\text{th}} = R_{\text{th}} + jX_{\text{th}} gives the whole curve, with its peak, the breakdown torque, at slip smax⁡=R2′/Rth2+(Xth+X2′)2s_{\max} = R_2'/\sqrt{R_{\text{th}}^2 + (X_{\text{th}} + X_2')^2}.

python
import numpy as np

# Per-phase equivalent circuit of an illustrative 400 V, 50 Hz, 4-pole induction motor (wye).
V_ph = 400 / np.sqrt(3)
f, poles = 50.0, 4
R1, X1 = 0.50, 1.00        # stator resistance and leakage reactance, ohm
R2, X2 = 0.60, 1.00        # rotor resistance and leakage reactance, referred to the stator, ohm
Xm = 28.0                  # magnetising reactance, ohm
n_sync = 120 * f / poles   # synchronous speed, rpm
w_sync = 2 * np.pi * n_sync / 60

# Thevenin equivalent of the supply and stator, seen from the rotor branch.
Z_th = (1j * Xm) * (R1 + 1j * X1) / (R1 + 1j * (X1 + Xm))
V_th = V_ph * (1j * Xm) / (R1 + 1j * (X1 + Xm))

def torque(s):
    """Electromagnetic torque (N m) at slip s: air-gap power 3 I2^2 R2/s divided by w_sync."""
    I2 = abs(V_th) / abs(Z_th + R2 / s + 1j * X2)
    return 3 * I2**2 * (R2 / s) / w_sync

def stator_current(s):
    Z = R1 + 1j * X1 + (1j * Xm) * (R2 / s + 1j * X2) / (R2 / s + 1j * (X2 + Xm))
    return abs(V_ph / Z)

s = np.linspace(1.0, 0.001, 2000)
T = np.array([torque(x) for x in s])
k = T.argmax()
s_max_formula = R2 / abs(Z_th.real + 1j * (Z_th.imag + X2))
print(f"synchronous speed {n_sync:.0f} rpm")
print(f"starting torque {T[0]:.0f} N m, starting current {stator_current(1.0):.0f} A")
print(f"breakdown (maximum) torque {T[k]:.0f} N m at slip {s[k]:.3f} ({n_sync * (1 - s[k]):.0f} rpm);"
      f" formula gives slip {s_max_formula:.3f}")

T_rated = 55.0                                       # the load the motor is meant to drive, N m
s_rated = s[np.argmin(np.abs(T[k:] - T_rated)) + k]  # on the stable side of the peak
n_rated = n_sync * (1 - s_rated)
print(f"at {T_rated:.0f} N m: slip {s_rated:.3f}, {n_rated:.0f} rpm, current {stator_current(s_rated):.1f} A,"
      f" output {T_rated * n_rated * 2 * np.pi / 60 / 1e3:.1f} kW")

print("\nspeed (rpm)  torque (N m)")
for n in range(0, 1501, 100):
    x = max(1 - n / n_sync, 1e-4)
    t = torque(x) if n < n_sync else 0.0
    print(f"{n:9d}   {t:8.1f}  " + "#" * int(round(t / 5)))
Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

This 8.3 kW motor runs its 55 N·m load at a slip of 3.7%, 1444 rpm, drawing 15.9 A. Started directly on the line it produces 113 N·m, about twice its rated torque, which is good, but draws 103 A, six and a half times its running current, which is why large motors are started through a soft-starter or a drive. The torque rises to a peak of 190 N·m at a slip of 0.296; on the steep right-hand side of that peak, more load means more slip and more torque, so the motor finds a stable operating point. Load it past the peak and it slows, loses torque and stalls.

starting torquebreakdown torqueoperating pointload torquesynchronousspeed (rpm)05001000150050100150200torque (N m)
The torque–speed curve of the induction motor simulated above, with a constant 55 N·m load (dashed). The motor starts with about twice its rated torque, passes through the breakdown torque, and settles where its curve crosses the load line, at a slip of a few per cent below the synchronous speed of 1500 rpm.

Induction motors are rugged, cheap, and need no magnets and no brushes, which is why they are the workhorses of industry. Their weaknesses are the magnetising current, which gives them a lagging power factor (the reason for the capacitor banks of the previous lesson), the rotor losses, and the fact that on a fixed-frequency supply they run at an almost fixed speed.

IntuitionH

Synchronous machines and generators

Replace the squirrel cage with a rotor that is itself a magnet, either a winding fed with DC (the field winding, through slip rings or a brushless exciter) or permanent magnets, and the rotor locks onto the rotating field and turns at exactly synchronous speed. That is a synchronous machine, and it works both ways: driven by a load it is a motor, driven by a turbine it is a generator. Nearly every large generator in the world is one: two-pole turbo-generators at 3000 or 3600 rpm in thermal power stations, and slow hydro generators with dozens of poles (a hydro generator with 80 poles turns at 75 rpm on a 50 Hz grid).

Per phase, a round-rotor synchronous generator is a voltage source E\mathbf E (its internal EMF, set by the field current) behind a synchronous reactance XsX_s. Connected to a grid of voltage V\mathbf V, with E\mathbf E leading V\mathbf V by the power angle δ\delta, it delivers per phase P=EVXssin⁡δ,Q=EVcos⁡δ−V2Xs.P = \frac{EV}{X_s}\sin\delta, \qquad Q = \frac{EV\cos\delta - V^2}{X_s}. Two separate controls follow. Real power is set by the turbine: open the steam or water valve, the rotor advances a few degrees relative to the grid, δ\delta grows and so does PP. Reactive power is set by the field current: raise EE above VV (over-excitation) and the machine supplies reactive power, holding up the grid's voltage; lower it and the machine absorbs reactive power. The power cannot exceed EV/XsEV/X_s, reached at δ=90∘\delta = 90^\circ; push harder and the rotor slips a pole and loses synchronism, which protection must trip at once.

Example— A generator's operating point, in per-unit

A generator with Xs=1.2X_s = 1.2 pu is connected to a 1.0 pu grid with its field set for E=1.6E = 1.6 pu. To deliver P=0.8P = 0.8 pu, sin⁡δ=0.8×1.2/1.6=0.6\sin\delta = 0.8 \times 1.2/1.6 = 0.6, so δ=36.9∘\delta = 36.9^\circ. Its reactive output is Q=(1.6×0.8−1)/1.2=0.233Q = (1.6 \times 0.8 - 1)/1.2 = 0.233 pu: over-excited, it supplies vars. Its limit at this excitation is Pmax⁡=1.6/1.2=1.33P_{\max} = 1.6/1.2 = 1.33 pu, so it runs with a margin before losing synchronism.

The same equations explain the stabilising inertia of the grid. The rotors of every synchronous generator in an interconnected system are locked to one frequency, and their stored kinetic energy cushions any sudden imbalance between supply and demand. The power grid builds the swing equation from this.

ExplanationH

Motors for electric vehicles and drones

Traction motors for vehicles must be light, efficient over a wide range of speeds and torques, and able to deliver high torque from standstill. Driven by an inverter, a motor's torque–speed envelope has two regions: up to the base speed, the inverter can push full current and the motor delivers its maximum torque; above it, the back-EMF approaches the battery voltage, the inverter weakens the motor's field to keep going, and the motor delivers roughly constant power, so the available torque falls as 1/ω1/\omega.

constant torqueconstant power (field weakening)torque ∝ 1/speedpowerbase speedmaximummotor speedtorque, power
The typical torque–speed envelope of an inverter-fed traction motor. Below the base speed the torque is limited by current and is constant (solid); above it, field weakening holds the power roughly constant (dashed), so the torque falls in inverse proportion to speed.

Four kinds of machine compete for this job in 2026:

  • Permanent-magnet synchronous motors (PMSM), especially with the magnets buried inside the rotor steel (interior PM), which adds a reluctance torque and makes field weakening easier. They are the most efficient and compact choice and power most battery-electric cars. Their magnets are neodymium–iron–boron, often with dysprosium or terbium added for high-temperature strength, and most of the world's rare-earth refining and magnet manufacture is in China (see the IEA's Global Critical Minerals Outlook 2025). Export restrictions announced in 2025 disrupted some car makers' magnet supplies and strengthened interest in the alternatives.
  • Induction motors, magnet-free and robust, with no back-EMF to worry about when coasting at high speed. Tesla's early Model S used them; several makers use one on a secondary axle that is often idle.
  • Wound-rotor (electrically excited) synchronous motors, with a field winding instead of magnets, so the field can be weakened at will. Renault and BMW are among the makers that have used them in production cars.
  • Switched-reluctance motors, with a rotor of plain toothed steel and no magnets or windings at all; the stator energises its coils in turn, and the rotor's teeth pull into line with whichever is on. They are cheap and tough but produce torque ripple and acoustic noise, and need their own kind of converter; they are used in some industrial drives and are an active research topic for magnet-free traction.

Drones use small permanent-magnet motors with the magnets on an outer rotor that spins round the stator (outrunners), driving propellers directly, each fed by its own electronic speed controller. Hobbyists rate them by KV, the no-load speed in rpm per volt, which is just the inverse of the back-EMF constant: k=60/(2π KV)k = 60/(2\pi\,\mathrm{KV}) V·s/rad. A 920 KV motor has k≈0.0104k \approx 0.0104 V·s/rad, so on a 14.8 V battery it would spin at about 13,600 rpm unloaded. The flight controller changes each motor's speed hundreds of times a second to steer, which you can follow up in Lumenary Robotics' aerial robots.

ExplanationH

Motor drives: a preview

A motor on a fixed 50 Hz supply runs at a nearly fixed speed. A variable-frequency drive (VFD) rectifies the supply to DC and uses an inverter to synthesise three-phase AC of any frequency and voltage, so it can run the motor at any speed with full torque. Two control methods dominate:

  • V/f control scales the voltage with the frequency to keep the flux constant (remember the EMF equation): simple, adequate for pumps and fans.
  • Field-oriented control (vector control) measures or estimates the rotor's position and splits the stator current into a part that makes flux and a part that makes torque, controlling each separately, so that an AC motor responds like the DC motor above, with torque proportional to one current. It is how EVs, robots and machine tools control their motors.

Drives save energy where the load varies. A fan's or pump's power rises roughly as the cube of its speed, so running a fan at 80% speed takes about 0.83≈51%0.8^3 \approx 51\% of full power, whereas throttling the flow with a damper at full speed saves far less. Efficiency is also regulated: the IEC defines efficiency classes from IE1 to IE5, and the European Union's Regulation 2019/1781 requires most three-phase motors from 0.75 to 1000 kW to meet at least IE3 (since July 2021), and those from 75 to 200 kW to meet IE4 (since July 2023). The switches, modulation and converters behind all this are the subject of Power electronics.

History & global context

A short history

Faraday demonstrated electromagnetic induction in 1831, and within a few years experimenters, among them Moritz Jacobi in St Petersburg and Thomas Davenport in Vermont, had built working DC motors. The transformer grew out of the "secondary generators" of Lucien Gaulard and John Dixon Gibbs (1882), and was made practical with a closed iron core by Károly Zipernowsky, Ottó Bláthy and Miksa Déri at the Ganz works in Budapest in 1884–85, and by William Stanley for Westinghouse in 1886. Galileo Ferraris and Nikola Tesla developed the induction motor from the rotating field in the late 1880s, and Mikhail Dolivo-Dobrovolsky gave it its squirrel-cage rotor. The rest of the story is materials and electronics: grain-oriented steel in the 1930s, rare-earth magnets (samarium–cobalt in the 1960s and 1970s, neodymium–iron–boron in 1984), and from the 1980s the power transistors that made variable-speed drives cheap.

Common misconceptionsH

Common misconceptions

Misconception

"A transformer makes power: step the voltage up and you get more." Power in equals power out, less losses. Stepping the voltage up by ten steps the current down by ten. That is the whole point for transmission: the same power at a tenth of the current has a hundredth of the I2RI^2R loss.

Misconception

"An induction motor runs at synchronous speed." It cannot: at synchronous speed the rotor sees no changing flux and makes no torque. A motor whose nameplate says 1450 rpm on a 50 Hz supply is a four-pole motor (1500 rpm synchronous) with 3.3% slip at full load.

Misconception

"A motor draws the most current when it is working hardest at full speed." It draws the most when it is stalled or starting, because there is no back-EMF to oppose the supply. A jammed motor can burn out its windings in seconds, which is why motors are protected by overload relays or fuses, and drives limit current.

Misconception

"Permanent-magnet motors are always better." They are usually the most efficient and compact, but magnets add cost and supply risk, lose strength when hot, and make back-EMF that the inverter must handle at high speed and in a fault. Induction, wound-rotor and reluctance machines each win in some applications.

Exercises

Exercises

1. Transformers and electric machines

Warm-upHuman

A transformer winding is to be connected to 230 V RMS at 50 Hz. Its core has a cross-section of 8 cm², and the peak flux density must be 1.5 T. Using V=4.44 fNΦmax⁡V = 4.44\,fN\Phi_{\max}, how many turns does the winding need, to the nearest whole turn?

2. Transformers and electric machines

Warm-upHuman

A six-pole induction motor runs on a 60 Hz supply with a full-load slip of 2.5%. What is its full-load speed in rpm?

3. Transformers and electric machines

CoreHuman

Why does a squirrel-cage induction motor produce no torque if its rotor turns at exactly synchronous speed?

Choose one

4. Transformers and electric machines

CoreHuman

A synchronous generator with synchronous reactance Xs=1.2X_s = 1.2 pu is connected to an infinite bus at V=1.0V = 1.0 pu. Its excitation gives an internal EMF E=1.6E = 1.6 pu. What power angle δ\delta, to the nearest degree, does it need to deliver P=0.8P = 0.8 pu?

5. Transformers and electric machines

CoreHuman

Some electric-vehicle makers use induction motors or wound-rotor synchronous motors instead of permanent-magnet motors. Which of these are genuine reasons? Select all that apply.

Choose all that apply

6. Transformers and electric machines

CoreComputational

Write operating_point(V, R, k, T_load, b=0.0) for a DC motor on a constant voltage: return (omega, current), the steady speed in rad/s and the armature current in A, from v=Ri+kωv = Ri + k\omega and ki=Tload+bωki = T_{\text{load}} + b\omega. If the stall torque kV/RkV/R cannot overcome the load, the motor stays stalled: return (0.0, V / R). The tests include the motor simulated in the lesson (it should give about 3331 rpm and 5.12 A) and an energy balance.

Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

7. Transformers and electric machines

CoreComputational

Write efficiency(x, pf, S_rated, P_core, P_cu_full), the efficiency of a transformer at load fraction x (of its rating S_rated in VA) and power factor pf, given its core loss and its copper loss at full load (in W). Then write best_load(P_core, P_cu_full), the load fraction at which the efficiency is highest. The tests use the lesson's 400 kVA transformer and a 1000 kVA one with 1.2 kW of core loss and 7.5 kW of full-load copper loss.

Runs entirely in your browser, in an isolated worker. Nothing is sent to a server.

Further reading and resources