IPM Motor Control: FOC Tuning, Wiring, and Sizing Guide

Interior Permanent Magnet (IPM) motor control relies on Field Oriented Control (FOC) to simultaneously manage magnetic torque and reluctance torque. Unlike surface-mounted magnet motors, an IPM rotor buries its magnets inside the laminations, creating a saliency ratio ($L_q > L_d$) that allows the drive to extract extra torque and extend the speed range via flux weakening. If you are designing a drive for an EV traction system, a high-speed HVAC compressor, or an industrial servo, IPM is the topology of choice for maximizing power density and efficiency.

This guide breaks down the exact controller requirements, terminal wiring, thermal sizing math, and failure signatures you need to deploy an IPM drive successfully on the bench or in the field.

Why IPM Motor Control Dominates High-Efficiency Drives

When matching a motor to a load profile, the deciding factor is often the constant power speed range (CPSR). Induction motors and Surface Permanent Magnet (SPM) motors struggle to maintain efficiency once you push past their base speed. IPM motors, however, thrive in variable-speed, high-inertia loads.

Because the magnets are buried, the d-axis inductance ($L_d$) is lower than the q-axis inductance ($L_q$). The FOC algorithm exploits this difference. By injecting negative d-axis current ($I_d$), the controller intentionally opposes the magnet's flux. This 'flux weakening' prevents the back-EMF from exceeding the DC bus voltage limit, allowing the motor to spin at 3x to 5x its base speed while maintaining constant power. This makes IPM the undisputed choice for loads that demand high starting torque and a wide operating speed range, such as electric vehicle powertrains and centrifugal compressors.

IPM vs. SPM vs. Induction: Motor Type Comparison

Selecting the right motor requires looking past the nameplate horsepower and evaluating the torque curve and control overhead. Below is a direct comparison of the three dominant AC motor topologies used in modern variable frequency drives (VFDs) and servo systems.

Wiring, Terminals, and Drive Requirements

An IPM motor will not run on a standard scalar V/f drive. It demands a vector drive capable of executing Space Vector PWM (SVPWM) at high switching frequencies (typically 10 kHz to 20 kHz) to minimize current ripple and acoustic noise. Controllers like the Texas Instruments C2000 series or dedicated Yaskawa servo amplifiers are standard for this task.

Power and Feedback Terminal Identification

When terminating an IPM motor, you are dealing with three distinct circuit domains. Miswiring the feedback domain will instantly fault the drive or cause a violent rotor jerk on startup.

• Power Terminals (U, V, W): The three-phase stator windings. These carry the high-frequency PWM waveforms. Use shielded, symmetrical motor cable (e.g., 4-conductor with a continuous corrugated aluminum armor acting as the ground/PE) to prevent common-mode bearing currents.
• DC Bus (+, -): If you are building the inverter, the DC link capacitors must be sized to handle the regenerative braking energy, as IPM motors act as highly efficient generators during deceleration.
• Resolver Feedback (6 wires): Most industrial IPMs use a brushless resolver rather than optical encoders due to vibration and thermal resilience. The terminals are typically labeled EXC+, EXC- (Excitation, usually 4-10 kHz AC sine wave), SIN+, SIN-, and COS+, COS-. The drive's Resolver-to-Digital (RDC) converter tracks these to calculate the exact electrical angle.
• Thermal Protection (T1, T2): Embedded PTC thermistors or PT1000 RTDs in the stator slots. IPMs are highly susceptible to demagnetization if the rotor exceeds the magnet's Curie temperature (often around 150°C for NdFeB grades).

Sizing Rule of Thumb and Worked Load Example

A common mistake is sizing an IPM motor and drive strictly on the continuous kW rating of the load. Because IPM control relies on injecting negative $I_d$ current for flux weakening at high speeds, the total RMS current in the stator ($I_s = \sqrt{I_d^2 + I_q^2}$) increases even when the mechanical torque output remains constant. Therefore, you must size the drive's continuous current rating to handle the thermal load of the combined current vector, not just the torque-producing $I_q$ component.

The Sizing Rule of Thumb: For applications requiring a 2:1 or greater constant power speed range, size the motor's continuous current rating and the drive's RMS output current at least 20% to 30% higher than the base-speed continuous load requirement.

Worked Example: 5 kW Centrifugal Compressor

Assume we are driving a 5 kW (approx. 6.7 HP) HVAC centrifugal compressor. The load requires 16 Nm of continuous torque at a base speed of 3,000 RPM, but must spin up to 9,000 RPM (a 3:1 speed range) while maintaining 5 kW output power.

1. Base Speed Current ($I_q$): At 3,000 RPM, the motor produces 16 Nm. Assuming a torque constant ($K_t$) of 1.2 Nm/A, the required q-axis current is $16 / 1.2 = 13.3A$ RMS.
2. High Speed Current ($I_s$): At 9,000 RPM, the back-EMF exceeds the 400V DC bus limit. The controller injects $-8A$ of $I_d$ current to weaken the flux. The $I_q$ required to maintain 5 kW at 9,000 RPM (which requires only 5.3 Nm of torque) drops to roughly 4.4A.
3. Total RMS Current at High Speed: $I_s = \sqrt{(-8)^2 + (4.4)^2} = \sqrt{64 + 19.36} = 9.1A$.

In this specific profile, the peak thermal current actually occurs at base speed (13.3A). However, if the mechanical load demanded high torque *and* high speed simultaneously, the $I_d$ and $I_q$ vectors would combine to push the RMS current well past the base-speed rating. Always verify the drive's continuous current rating (e.g., a 20A rated Yaskawa Sigma-7 servo amplifier) against the worst-case vector sum across the entire torque-speed envelope.

Failure Signatures: Diagnosing Hum, Overheat, and Stall

IPM drives do not fail silently. The FOC algorithm's reliance on precise mathematical models of the motor's inductance and rotor position means that physical degradation or tuning errors manifest in distinct, measurable ways.

1. Acoustic Hum and Cogging (The Angle Offset Error)

Symptom: The motor emits a loud, high-pitch whine or low-frequency cogging vibration at low speeds, and draws excessive no-load current.
Cause: Incorrect resolver offset angle or cross-coupling in the PI current loops. If the electrical angle is off by even 5 to 10 degrees, the Park/Clarke transforms misalign. The controller thinks it is injecting pure torque-producing current ($I_q$), but a portion of it is actually being injected into the d-axis ($I_d$), fighting the magnets.
Fix: Run the drive's static auto-tune routine. If the hum persists, manually verify the pole-pair count parameter in the drive. A mismatch between the physical pole pairs (e.g., 4) and the drive parameter (e.g., 8) will cause continuous electrical angle drift.

2. Stator Overheat at High Speed (Flux Weakening Saturation)

Symptom: The motor casing is cool at base speed but becomes dangerously hot (triggering the PTC thermistor fault) when operating above base speed, even under light mechanical loads.
Cause: Excessive $I_d$ injection. If the motor's actual $L_d$ inductance is lower than the value programmed into the FOC algorithm (often due to magnetic saturation at high currents), the controller will inject more negative $I_d$ current than necessary to weaken the flux. This wasted current generates $I^2R$ heat in the stator windings without producing torque.
Fix: Reduce the flux-weakening gain in the drive parameters, or perform an offline inductance saturation curve measurement to update the drive's motor model lookup tables.

3. High-Speed Stall and Desync (PLL Tracking Failure)

Symptom: The motor runs smoothly up to 6,000 RPM, then suddenly faults with an 'Overcurrent' or 'Position Loss' error, violently braking the load.
Cause: The Phase-Locked Loop (PLL) inside the Resolver-to-Digital converter cannot track the rapid rate of change of the rotor angle. At high RPM, the electrical frequency can exceed 1 kHz. If the resolver excitation frequency is too low (e.g., 2 kHz) or the PLL bandwidth is tuned too narrow, the drive loses track of the rotor position, causing the FOC algorithm to commutate the wrong phases.
Fix: Increase the resolver excitation frequency to at least 10 kHz (ensuring the resolver hardware supports it) and widen the PLL tracking bandwidth in the drive's advanced configuration menu.

IPM Motor Control FAQ

How does flux weakening work in IPM motor control?

Flux weakening allows an IPM motor to operate above its base speed without exceeding the voltage limit of the DC bus. As the rotor spins faster, the permanent magnets generate a higher back-EMF voltage. Once this back-EMF approaches the DC bus voltage, the motor can no longer draw current to produce torque. The FOC controller solves this by intentionally injecting negative current along the d-axis ($I_d$). This current creates a magnetic field that directly opposes and 'cancels out' a portion of the permanent magnet's flux. By weakening the total magnetic flux in the air gap, the back-EMF drops, allowing the motor to spin faster while maintaining constant power output, albeit with reduced torque capability.

Why does my IPM drive require a resolver instead of hall sensors?

Hall sensors provide discrete, low-resolution position updates (typically 6 steps per electrical revolution). While this is sufficient for basic trapezoidal commutation in SPM motors, IPM motor control relies on Field Oriented Control (FOC), which requires continuous, high-resolution knowledge of the exact rotor electrical angle to calculate the Park and Clarke transforms accurately. A resolver provides an absolute, continuous analog sine/cosine signal that the drive's RDC (Resolver-to-Digital) chip converts into high-resolution digital angles (often 16-bit or higher). Furthermore, resolvers are transformer-based and lack semiconductor components, making them immune to the high electromagnetic interference (EMI) and thermal extremes generated by high-frequency SiC/IGBT inverters.

What is the difference between IPM and SPM motor control algorithms?

The core difference lies in how the algorithms handle the d-axis current ($I_d$). In a Surface Permanent Magnet (SPM) motor, the magnets are on the surface, meaning the rotor is non-salient ($L_d = L_q$). The control algorithm simply forces $I_d$ to zero and uses all available current for torque-producing $I_q$. In an IPM motor, the buried magnets create saliency ($L_q > L_d$). The IPM control algorithm must use a Max Torque Per Ampere (MTPA) trajectory, intentionally injecting both $I_q$ and a specific amount of positive or negative $I_d$ to exploit reluctance torque. Additionally, the IPM algorithm requires a complex flux-weakening voltage limit loop that the SPM algorithm does not need, making the IPM control code significantly more computationally demanding on the microcontroller.