- Author: Johnny Liu, CEO at Dowway Vehicle
- Published: June 2026
- Target Audience: EV Powertrain Engineers, Motor Control Software Developers, VCU Calibration Engineers
- Reading Time: 12 minutes
📌 Executive Summary (TL;DR)
What is PMSM Stall Control? Under a locked-rotor (stall) condition, a Permanent Magnet Synchronous Motor (PMSM) has zero rotor speed ($n \approx 0$), resulting in zero mechanical output power ($P_{mech} \approx 0$). However, the electrical system remains highly active. Instead of “increasing current to push through,” a robust motor controller must dynamically arbitrate torque commands against a strict cooperative system of limiters (current, thermal, voltage, and demagnetization boundaries). Effective stall control balances immediate vehicle drivability (escape capability) with long-term hardware survivability through a multi-layer FOC framework and VCU-MCU coordination.
Table of Contents
Introduction: The Paradox of Zero Speed and High Power
In the field of New Energy Vehicles (NEVs) and electric drive systems, the Permanent Magnet Synchronous Motor (PMSM) is widely praised for its high efficiency, high power density, and rapid dynamic response. However, these performance advantages come with a major engineering caveat: the motor control unit (MCU) must maintain absolute precision during extreme corner cases.
Among these corner cases, the stall (locked-rotor) condition is one of the most physically punishing.
[Vehicle Control Unit (VCU)]
│ (Torque Request)
▼
[Motor Control Unit (MCU)] ──(High Current Injection)──> [Inverter & PMSM (Stalled)]
│
┌─────────────────────┴─────────────────────┐
▼ ▼
[Zero Mechanical Output] [High Thermal Loss]
P_mech = Te * ω_m ≈ 0 P_cu ≈ I_rms^2 * R_s (Heat!)
Figure 1: Energy flow and thermal conversion pathways in a locked-rotor PMSM system.
By definition, a stall occurs when the controller commands electromagnetic torque, but the rotor remains stationary or rotates at a speed near zero due to external factors. Common real-world causes include:
- Extreme uphill starts under heavy loads
- Tires wedged in deep mud, snow, or off-road obstacles
- Mechanical transmission jams or seized bearings
- Unreleased mechanical parking brakes or service brakes
When a vehicle fails to move despite the driver stepping on the accelerator, a common misconception is that the motor is “not working.” In reality, the exact opposite is true. Because the rotor is stationary, the mechanical output power of the system is mathematically zero:$$P_{mech} = T_e \cdot \omega_m \approx 0$$
Where:
- $T_e$ is the electromagnetic torque ($\text{N}\cdot\text{m}$)
- $\omega_m$ is the mechanical angular velocity ($\text{rad/s}$)
Since mechanical power output is zero, conservation of energy dictates that virtually all electrical energy injected into the stator winding is converted directly into thermal energy. This is dominated by stator winding copper loss ($P_{cu}$) and inverter switching/conduction losses:$$P_{cu} \approx I_{rms}^2 R_s$$
Where:
- $I_{rms}$ is the Root-Mean-Square phase current ($\text{A}$)
- $R_s$ is the stator winding resistance ($\Omega$)
Without effective intervention, this localized heat generation can lead to catastrophic system failure within seconds. Therefore, the core objective of PMSM Stall Control is not to blindly increase current to overcome the obstacle, but to find the precise equilibrium between maximizing torque output and protecting the electrical drive hardware.
1. Back to Basics: Torque Generation in FOC Coordinate Systems
To understand how a controller manages a stalled motor, we must look at how electromagnetic torque is mathematically generated under Field-Oriented Control (FOC).
FOC decouples the three-phase stator currents ($i_a$, $i_b$, $i_c$) into a two-axis rotating coordinate system aligned with the rotor’s magnetic flux (the $d/q$ reference frame) using Clarke and Park transformations:
- $d$-axis current ($i_d$): Aligned with the rotor permanent magnet flux linkage direction. It regulates excitation, field-weakening (at high speeds), and influences reluctance torque in salient-pole motors.
- $q$-axis current ($i_q$): Orthogonal to the rotor flux linkage direction. It interacts directly with the magnet flux to generate the primary electromagnetic torque.
q-axis (Torque Generation)
▲
│ / Rotor Position (θ_e)
│ /
│ /
│ /
│ /
│ /
│ /
────────────────────┼───────────────────► d-axis (Flux Linkage / Field Weakening)
/│
/ │
/ │
Figure 2: Geometric alignment of d/q coordinate axes relative to the rotor magnetic pole.
The fundamental electromagnetic torque equation for a three-phase PMSM is:$$T_e = \frac{3}{2}p \left[ \psi_f i_q + (L_d – L_q) i_d i_q \right]$$
Where:
- $p$ is the motor pole pairs
- $\psi_f$ is the permanent magnet flux linkage ($\text{Wb}$)
- $L_d, L_q$ are the $d$-axis and $q$-axis inductances ($\text{H}$), respectively
- $i_d, i_q$ are the $d$-axis and $q$-axis stator currents ($\text{A}$), respectively
SPMSM vs. IPMSM Torque Characteristics
- Surface PMSM (SPMSM): Since the permanent magnets are mounted on the rotor surface, the magnetic paths are symmetrical, resulting in $L_d \approx L_q$. The torque equation simplifies to: $$T_e = \frac{3}{2}p \psi_f i_q$$ For an SPMSM, the controller typically executes an $i_d = 0$ control strategy, making torque output strictly proportional to $i_q$.
- Interior PMSM (IPMSM): Because the magnets are embedded inside the rotor core, there is structural saliency where $L_q > L_d$. This produces negative reluctance torque. To maximize torque output per unit of current during heavy load or stall startup, the MCU utilizes Maximum Torque Per Ampere (MTPA) algorithms to dynamically inject negative $i_d$ alongside $i_q$, leveraging both magnetic alignment and reluctance forces.
Under normal driving, the Vehicle Control Unit (VCU) sends a torque request ($T_{req}$) to the motor controller. The MCU converts this request into target currents ($i_{d,ref}, i_{q,ref}$), which are regulated by the FOC current loop.
[VCU Torque Request (T_req)] ──► [MCU Boundary Limiters] ──► [Torque Command (T_cmd)] ──► [MTPA Solver] ──► [Target Currents (id,ref / iq,ref)] ──► [FOC Current Loop]
Figure 3: Signal transmission path from vehicle-level torque request to FOC current regulation.
2. Zero-Speed FOC Challenge: Sensored vs. Sensorless Position Estimation
To execute FOC successfully, the controller must know the precise electrical rotor angle ($\theta_e$). If the calculated angle is incorrect, the coordinate transformation will misalign: the intended $q$-axis current will leak into the $d$-axis. This reduces torque output, increases thermal losses, and can even cause motor reversal.
In zero-speed stall conditions, acquiring this angle presents vastly different challenges depending on the system architecture.
Comparative Analysis: Sensored vs. Sensorless Zero-Speed FOC
| Feature / Metric | Sensored FOC (Resolver, Encoder, or Hall) | Sensorless FOC (Observer / Sliding Mode / Back-EMF) |
|---|---|---|
| Angle Source at Zero Speed | Direct physical measurement (accurate down to 0 RPM). | No physical sensor; relies on software estimation models. |
| Back-EMF Amplitude | Irrelevant to angle feedback quality. | Approaching zero ($e \propto \omega_m \approx 0$). Signal-to-noise ratio is extremely low. |
| Torque Control Capability | High and continuous; can maintain rated stall torque indefinitely (subject to thermal limits). | Extremely low under standard observer models due to loss of angle observability. |
| Mitigation Strategies | Regular current regulation and thermal monitoring. | Rotor pre-positioning (DC injection), open-loop forced pulling (I-f control), or High-Frequency Injection (HFI). |
| Risk Profile & Margin | Low risk of angle loss; thermal dissipation is the primary concern. | High risk of angle drift or phase step-out. Requires conservative torque limits and rapid fault triggers. |
Engineering Implications
For heavy-duty commercial vehicles, off-road machinery, and mining trucks, sensored feedback (typically via high-resolution resolvers) is mandatory. It ensures the MCU can deliver high torque at zero speed to prevent vehicle rollback on steep slopes.
In contrast, consumer or industrial applications using sensorless FOC must rely on advanced saliency-tracking methods like High-Frequency Injection (HFI) or resort to strict low-speed torque clamping to prevent the motor from slipping out of control during a stall.
3. Beyond the PI Loop: Stall Control as a System of Limiters
A common engineering mistake is assuming that a standard speed-torque closed-loop controller can handle a stall.
If a vehicle is stuck, a speed PI controller will see a massive error between the target speed (e.g., 100 RPM) and the actual speed (0 RPM). The speed loop’s integral term will rapidly saturate, commanding maximum torque and pushing the current loop to its absolute physical limits. Without protective limiters, this will quickly cause thermal runaway in the inverter or demagnetize the rotor.
Robust stall control must run parallel to the PI loops, operating as a cooperative system of limiters that continuously calculate the maximum safe torque envelope.
┌─────────────────────────────────────────┐
│ - Maximum Phase Current Limit │
│ - q-Axis Current Limit │
│ - Motor Winding Temperature (Temp_mot) │ ──► Min() Selector ──► Current Limit Envelope (I_max)
│ - Inverter Junction Temp (Temp_igbt) │
│ - DC-Link Voltage / Current Limits │
│ - Rotor Demagnetization Limit │
└─────────────────────────────────────────┘
│
▼
[VCU Torque Request (T_req)] ─────────────────► [Limiter] ─────────────────► [Allowed System Torque (T_avail)]
Figure 4: Dynamic arbitration of the torque envelope based on cooperative hardware limiters.
The actual torque output ($T_{avail}$) is the minimum value allowed by all active system limits:$$T_{avail} = \min \left( T_{req}, T_{limits} \right)$$
These limits include:
- Maximum Phase Current Limit ($I_{max,phase}$): Restricts the peak current flowing through the stator windings to prevent instantaneous overcurrent faults.
- $i_q$ Current Limit ($I_{q,limit}$): Prevents excessive torque ripple and protects mechanical transmission shafts from shearing under sudden stress.
- Motor Winding Temperature Limits ($T_{winding}$): Monitored via physical NTC/PTC sensors or calculated using real-time thermal observers. When temperatures exceed calibrated thresholds (e.g., $140^\circ\text{C}$), the controller initiates thermal derating.
- Inverter IGBT/MOSFET Thermal Limits ($T_{junction}$): Power semiconductor junctions have very small thermal time constants (milliseconds). The MCU must monitor or estimate junction temperatures and apply rapid derating to prevent thermal breakdown of the silicon.
- DC-Link Voltage & Current Limits ($V_{dc}, I_{dc}$): Keeps the battery discharge power within limits, protecting the BMS and preventing excessive voltage drops on the DC bus.
- Demagnetization Risk Limit: Neodymium-Iron-Boron (NdFeB) permanent magnets in PMSMs suffer irreversible demagnetization if subjected to high temperatures combined with a strong opposing magnetic field (large negative $i_d$ current). The controller must clamp $i_d$ and limit temperatures under stall conditions to avoid ruining the rotor.
4. Multi-Variable Stall Detection and Protection State Machines
To protect the drive system without triggering false alarms during brief heavy-load starts, the MCU must use multi-variable detection logic coupled with a phased protection state machine.
Stall Detection Logic
The MCU flags a stall condition when the following conditions are simultaneously met for a calibrated duration ($t > t_{threshold}$):$$\text{Stall} = \left( |T_{req}| \ge T_{thresh} \right) \;\wedge\; \left( |n_{rotor}| \le n_{thresh} \right) \;\wedge\; \left( I_{rms} \ge I_{thresh} \right) \;\wedge\; \left( t \ge t_{stall\_delay} \right)$$
To improve detection accuracy and prevent false positives (such as high-acceleration launches), the MCU cross-checks other vehicle parameters via the CAN/CAN-FD bus, including:
- Brake pedal status (active brake-torque fighting)
- Selected gear (D/R vs. N/P)
- Individual wheel speed sensors (detecting differential wheel slip vs. true vehicle immobility)
Phased Protection State Machine
Once a stall is confirmed, the MCU transitions through a phased protection state machine to safely manage the event.
┌─────────────────┐ Stall Detected ┌───────────────────┐ Temp/Time Rises ┌───────────────────┐
│ │ ──────────────────────────► │ Phase 1: │ ──────────────────────────► │ Phase 2: │
│ Normal Drive │ │ Torque Clamping │ │ Thermal Derating │
│ │ ◄────────────────────────── │ (iq reduction) │ ◄────────────────────────── │ (Dynamic limits) │
└─────────────────┘ Stall Recovered └───────────────────┘ Stall Recovered └───────────────────┘
▲ │
│ │ Temp > Max Limit
│ ▼
┌─────────────────┐ ┌───────────────────┐
│ │ ◄──────────────────────────────────────────────────────────────────────────── │ Phase 3: │
│ Drive Disable │ │ Fault & Warning │
│ (Safe State) │ ◄─────────────────────────────────────── UDS Fault Active ─────────────────── │ (UDS DTC Active) │
└─────────────────┘ └───────────────────┘
Figure 5: Phased state machine transitions during a persistent motor stall event.
- Phase 1: Torque Clamping (Immediate). The controller limits the maximum $i_q$ to a temporary safe value (e.g., 70% of peak capability). This allows the driver a brief window to climb over an obstacle or roll back safely.
- Phase 2: Thermal Derating (Progressive). If the stall persists, the controller monitors stator and IGBT temperatures. The allowed torque is scaled down dynamically along a pre-defined thermal slope, reducing current to the continuous rating of the electric drive.
- Phase 3: Fault Warning & Diagnostics (UDS DTC Reporting). If winding temperatures continue to climb towards critical thresholds, the MCU broadcasts a high-priority warning frame over the CAN bus, illuminates the dashboard malfunction indicator light (MIL), and logs a Diagnostic Trouble Code (DTC) via UDS (Unified Diagnostic Services).
- Phase 4: Drive Disable & Safe State (Emergency Shutdown). If the temperature exceeds the maximum safe threshold (e.g., $180^\circ\text{C}$ winding temperature or $150^\circ\text{C}$ IGBT junction temperature), the controller disables the inverter switches. The system transitions to a “Safe State” (e.g., Active Short Circuit – ASC, or Free Wheel – Safe Pulse Off) to protect the hardware from thermal destruction.
Hysteresis and Recovery
To prevent “hunting” (where the system rapidly enters and exits protection, causing severe driveline oscillations), the recovery path must include hysteresis. The system will only exit protection and restore normal torque limits once the stall condition clears and temperatures fall below a safe recovery threshold (e.g., $T_{winding} < 110^\circ\text{C}$).
5. System-Level Coordination: VCU Calibration and Integration
Stall control is not just an isolated motor control problem; it requires careful coordination with the vehicle control unit (VCU).
If the VCU torque ramp rate is too steep, or if the driver’s pedal requests are unfiltered during a stall, the MCU current loop will experience massive transients. This can trigger early thermal protection and stall out the vehicle unnecessarily.
VCU Stall Calibration Parameters and Strategies
| Calibration Target | Primary System Risks | Recommended Mitigation & Calibration Strategies |
|---|---|---|
| Launch Torque Request ($T_{launch}$) | * Instantaneous overcurrent trips.* Abrupt drivetrain jerking.* Premature thermal derating. | * Implement split torque maps for low-speed and stall conditions.* Map torque requests to vehicle speed ($n$) and apply active slope compensation. |
| Pedal Ramp Rate / Slope ($dT/dt$) | * Mechanical driveline lash/backlash.* High current spikes in the MCU.* Traction loss in low-mu surfaces. | * Calibrate low-pass filters on pedal inputs.* Set dynamic rate limiters based on current speed ($n \approx 0$). Allow a temporary torque window for obstacle escape before clamping. |
| Brake-Release Coordination | * “Brake-torque fighting” (wasting energy as heat).* Accelerated mechanical brake wear.* Winding hot spots. | * Implement cooperative arbitration: when both brake and accelerator pedals are pressed, prioritize brake signals.* Verify gear position, wheel speed, and EPB status before ramping up driving torque. |
| Thermal Recovery Hysteresis | * Oscillating torque delivery.* Winding insulation fatigue from thermal cycling.* Unpredictable vehicle acceleration. | * Introduce temperature hysteresis thresholds (e.g., $\Delta T \ge 25^\circ\text{C}$).* Incorporate time-delay locks to prevent repeated, high-frequency stall escape attempts. |
6. The 5-Layer Engineering Framework for Robust Stall Control
To build a reliable electric drive system, stall control should be organized into a five-layer software architecture. This structure separates high-level vehicle demands from low-level hardware safety overrides.
┌────────────────────────────────────────────────────────────────────────┐
│ Layer 1: Torque Request Layer (Driver & VCU Arbitration) │
└────────────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Layer 2: Current Execution Layer (FOC Current Loops & MTPA) │
└────────────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Layer 3: State Identification Layer (Speed, Current & Time Observers) │
└────────────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Layer 4: Capability Limitation Layer (Thermal & Voltage Models) │
└────────────────────────────────────────────────────────────────────────┘
│
▼
┌────────────────────────────────────────────────────────────────────────┐
│ Layer 5: Fault Protection Layer (Diagnostics, Safe State & UDS) │
└────────────────────────────────────────────────────────────────────────┘
Figure 6: The 5-layer engineering framework for comprehensive PMSM stall management.
Layer 1: Torque Request Layer
Responsible for receiving driver commands and VCU requests. It arbitrates between different vehicle functions (e.g., cruise control, traction control, slope hold) and passes a consolidated torque target ($T_{req}$) to the motor controller.
Layer 2: Current Execution Layer
The core of the FOC algorithm. It runs high-frequency current loops (typically $10\text{ kHz}$ to $20\text{ kHz}$) and uses MTPA lookup tables to convert torque commands into optimal $i_d$ and $i_q$ current references.
Layer 3: State Identification Layer
Continuously monitors system state variables. By checking resolver feedback, phase currents, and VCU status signals, it detects when the motor is stalled and manages the stall confirmation timers.
Layer 4: Capability Limitation Layer
The thermal and electrical safety engine. It calculates physical operating boundaries in real time based on active temperature sensors and dynamic thermal models. It scales back allowable current limits ($I_{max}$) before any physical components reach dangerous temperatures.
Layer 5: Fault Protection Layer
The final safety line. If the stall persists beyond safe thermal or time boundaries, Layer 5 overrides all lower layers. It logs diagnostic fault codes, alerts the vehicle coordinator, and disables the inverter to protect the physical hardware.
Conclusion: The Silent Battle of Electric Powertrains
A motor’s reliability is not just tested at $15,000\text{ RPM}$; it is defined at $0\text{ RPM}$.
PMSM stall control represents a balancing act between vehicle drivability and hardware protection. FOC algorithms ensure the motor produces torque efficiently, while the stall protection system determines how long that torque can be safely maintained before the hardware must step back.
For motor control engineers, mastering stall control requires a deep understanding of current loops, angle estimation, thermal modeling, and diagnostic state machines. For VCU calibration engineers, it requires balancing torque ramps, driving feel, and hardware limits.
When an electric vehicle effortlessly climbs a steep rocky path or pulls itself out of deep mud, it is easy to focus on the raw power. But behind the scenes, a silent battle is being won by the motor controller—proving that in electric powertrains, absolute control at zero speed is the true mark of engineering excellence.
FAQ: High-Value Technical Reference
Q1: Why does a PMSM heat up so rapidly during a stall if the vehicle is not moving?
At zero rotor speed, the mechanical output power ($P_{mech} = T_e \cdot \omega_m$) is zero. However, to maintain the requested torque, the motor controller must inject large currents into the stator windings. Without rotor movement, there is no back-EMF ($e \approx 0$), meaning the applied voltage goes entirely toward overcoming winding resistance ($R_s$). This converts all electrical energy into stator copper loss ($P_{cu} \approx I_{rms}^2 R_s$). Additionally, because the rotor is stationary, heat is concentrated in a single physical location rather than being distributed evenly across all phases, leading to rapid localized hot spots.
Q2: What is the difference between peak current and thermal derating current in motor controllers?
- Peak Current: The maximum current the inverter and motor can handle for a very short duration (typically 3 to 10 seconds). This limit is determined by the maximum junction temperature ($T_{j}$) of the IGBTs/MOSFETs and the magnetic saturation of the stator core.
- Thermal Derating Current: A dynamically adjusted current limit calculated by the MCU’s thermal management models. As stator winding or power semiconductor temperatures rise, the controller reduces the allowable current from the peak limit toward the motor’s continuous current rating to prevent thermal damage.
Q3: How does demagnetization occur during a stall event, and how can it be prevented?
Demagnetization occurs when permanent magnets (such as NdFeB) are exposed to high temperatures and a strong opposing magnetic field (demagnetizing field) simultaneously. In a stall condition, the combination of high stator copper loss (heat) and large negative $d$-axis currents (used in MTPA or field-weakening strategies) can push the magnets past their knee point on the $B\text{-}H$ magnetic hysteresis curve, causing irreversible loss of magnetic flux.
This is prevented by:
- Implementing dynamic, low-latency thermal observers for the rotor.
- Clamping the maximum negative $d$-axis current limit as rotor temperatures increase.
- Designing the rotor structure with protective steel bridges to shield magnets from direct stator flux transients.