EE REFRESHEREELESSONS // S1
EPISODE 03
Embedded & Control Systems

Field-Oriented Control of PMSMs

"How to make a three-phase motor think it's a DC motor."

⏱ ~14 minHosts: Elena & Marcus Read the transcript Companion notes (PDF)
Download mp3

In this episode

A brushed DC motor is trivial to control because the commutator holds field and armature current orthogonal, so torque is just proportional to current — one decoupled knob. A permanent-magnet synchronous motor is wonderful hardware but hostile in its native abc frame: three coupled, time-varying sinusoids and a spinning rotor. Field-oriented control fixes the frame, not the hardware. The Clarke transform collapses abc to a two-axis stationary αβ frame; the Park transform rotates that frame to spin with the rotor, into d-q — where the sinusoids become DC and a plain PI loop can regulate them with zero steady-state error. Torque ∝ iq; id handles flux. Then the modern layer: field weakening (negative id to push past base speed), MTPA and reluctance torque in interior PMSMs, and sensorless operation that deletes the shaft encoder.

Lesson map

  1. ObjectiveUnderstand vector control well enough to explain why it works.
  2. The refresherWhy the DC motor is easy (orthogonal, decoupled) and the 3-phase machine is hard in abc.
  3. What changedCheap fast MCUs + SiC inverters + mature sensorless observers made vector control universal.
  4. The mechanism — Clarke & Parkabc → αβ → dq; sinusoids become DC; two PI loops give decoupled torque/flux control.
  5. Going modernField weakening, MTPA & reluctance torque in IPMs, sensorless observers + HF injection.
  6. Industry standard + self-testEV traction & servo drives, turnkey libraries, two exercises.

Key concepts

Clarke transform (abc→αβ)Park transform (αβ→dq)d-q frame iq (torque) / id (flux)PI current loopsSpace-vector modulation Back-EMFField weakeningMTPAReluctance torque Interior PMSMSensorless / sliding-mode observerHF injectionPole pairs

Self-test

1 · Electrical vs. mechanical frequency

Electrical frequency = mechanical rotation frequency × pole pairs. For a motor with 4 pole pairs spinning at 3000 RPM, what's the electrical frequency your Park transform and current loop are tracking? (Answer lands in the hundreds of hertz — which is why the current loop runs at tens of kHz.) That loop-rate-to-electrical-frequency ratio is a real design constraint.

2 · Why id = 0 isn't always right

Explain why holding id = 0 is correct for a surface-magnet motor but leaves performance on the table for an interior PMSM. The key: buried magnets create d-q inductance asymmetry → reluctance torque ∝ id·iq → an optimal nonzero id exists (maximum torque per amp).

Warm up in CircuitQuEEst

The matching undergraduate foundations are quests in CircuitQuEEst, the free, offline Android app for learning EE (download the APK). Play these first, then come back for the graduate-level version:

Go deeper