🤖 Week 7 Summary: Actuators & Drive Systems
Theme: Actuators & Drive Systems
Topic: Weekly Summary and Key Takeaways
Learning Goal: Consolidate understanding of robotic actuation and connect to broader robotics knowledge.
What We Covered This Week
| Day | Topic | Core Concept |
|---|---|---|
| 1 | Actuator Overview | Three families (electric, hydraulic, pneumatic) and selection criteria |
| 2 | DC & Servo Motors | Motor physics, back-EMF, servo control loops (P/PI/PID) |
| 3 | Stepper & BLDC Motors | Open-loop precision, electronic commutation, FOC |
| 4 | Gearboxes | Harmonic drives (zero backlash), planetary gears, cycloidal drives |
| 5 | Motor Drivers | H-bridges, PWM, MOSFETs, protection circuits, power electronics |
| 6 | Python Practice | Motor simulation, PID tuning, torque-speed analysis |
Key Equations
Motor Physics
| Equation | Meaning | When to Use |
|---|---|---|
τ = Kt × I | Torque proportional to current | Calculating required current for given torque |
V = R×I + L(dI/dt) + Ke×ω | Complete motor electrical model | Transient analysis, driver design |
ω_no_load = V / Ke | Maximum speed at given voltage | Sizing motors for speed requirements |
τ_stall = Kt × V / R | Maximum torque (at zero speed) | Worst-case load analysis |
η = (τ×ω) / (V×I) | Motor efficiency | Energy budgeting, thermal design |
Gearbox Relationships
| Equation | Meaning | When to Use |
|---|---|---|
τ_out = τ_in × N × η | Output torque after gearbox | Load capacity verification |
ω_out = ω_in / N | Output speed reduction | Speed requirement verification |
J_reflected = J_load / N² | Inertia seen at motor shaft | Dynamic response analysis |
Control
| Equation | Meaning | When to Use |
|---|---|---|
u = Kp×e + Ki∫e + Kd(de/dt) | PID control law | Position/velocity control tuning |
Bandwidth ≈ Kp / (2π) | Approximate position loop bandwidth | Performance estimation |
D = T_on / (T_on + T_off) | PWM duty cycle | Motor speed/voltage control |
Common Mistakes
1. Ignoring Reflected Inertia
Wrong: Size motor only for steady-state torque.
Right: Include dynamic torque for acceleration:
τ_total = τ_friction + τ_gravity + J_total × α
Impact: Underpowered motors that can’t achieve required acceleration.
2. Confusing Holding Torque with Dynamic Torque
Wrong: Use stepper holding torque as operating torque.
Right: Torque drops with speed — use torque-speed curve at operating RPM.
Impact: Missed steps, position errors, stalled motors.
3. Neglecting Thermal Limits
Wrong: Run motor at peak torque continuously.
Right: Calculate RMS torque over motion profile; verify against rated continuous torque.
Impact: Overheated motors, insulation damage, reduced lifespan.
4. Forgetting Gearbox Efficiency
Wrong: τ_out = τ_in × N (no efficiency factor).
Right: τ_out = τ_in × N × η (typically η = 0.80-0.95).
Impact: Insufficient output torque, oversized motors.
5. Open-Loop Stepper Without Safety Margin
Wrong: Operate stepper at 90% of holding torque.
Right: Maintain 50% torque margin; add encoder for closed-loop.
Impact: Missed steps under load spikes, accumulated position error.
Connections to Previous Weeks
| Previous Week | Connection to Actuators |
|---|---|
| W1 (Robot Basics) | Actuators are the “muscles” that execute motion commands |
| W2 (Forward Kinematics) | Joint angles → actuator positions; gear ratios scale encoder resolution |
| W3 (Inverse Kinematics) | IK solutions must respect actuator position and velocity limits |
| W4 (Path Planning) | Trajectories must respect actuator acceleration and torque constraints |
| W5 (Dynamics) | τ = M(q)q̈ + C(q,q̇)q̇ + G(q) — dynamics tells us what torque each actuator must produce |
| W6 (Sensors) | Encoders (position feedback) close the control loop; current sensors enable torque control |
Key insight: Actuators are where all previous weeks converge. Kinematics tells us where to go, dynamics tells us what forces are needed, and actuators provide the capability to make it happen.
Hardware vs. Simulation
| Aspect | Simulation | Real Hardware |
|---|---|---|
| Speed | Instant | Limited by physics |
| Cost | Free | $50-5000 per joint |
| Safety | No risk | High current, moving parts |
| Accuracy | Perfect model | Friction, backlash, noise |
| Learning value | Theory | Real-world constraints |
Recommendation: Use simulation to develop intuition and test algorithms, but validate on hardware early. The gap between simulation and reality is where the most important engineering lessons live.
What to Watch Next
Week 8 Preview: ROS2 and System Integration
We’ll bring everything together:
- ROS2 node architecture: Perception (W6) + Planning (W4) + Control (W7)
- Joint trajectory execution: From planned path to actuator commands
- Hardware interfaces: ros2_control and real motor drivers
- Simulation in Gazebo: Validate before hardware deployment
The goal: A complete robot software stack that connects sensors → algorithms → actuators in a unified framework.
Resources and References
Essential Reading
- Siciliano & Khatib — Springer Handbook of Robotics, Chapters 5-6
- Spong, Hutchinson, Vidyasagar — Robot Modeling and Control, Chapter 6 (Actuators)
- Maxon Motor — Formulae Handbook (free download)
Online Resources
- ODrive: odriverobotics.com — Open-source FOC controller
- VESC: vedder.se — Open-source BLDC controller
- Harmonic Drive: harmonicdrive.net — Technical documentation
Datasheets to Study
- Maxon RE/EC motor series (DC and BLDC)
- Harmonic Drive CSF/SHF series
- Trinamic TMC2209 stepper driver
- TI DRV8301/DRV8323 BLDC gate drivers
Self-Test Questions
-
Why does a stepper motor lose torque at high speeds?
Inductance limits current rise rate:
di/dt = (V - Ke×ω) / L. At high speed, back-EMF reduces effective voltage. -
What’s the advantage of harmonic drives over planetary gears for robot joints?
Zero backlash and compact size. Backlash causes position error when reversing direction, critical for precision robots.
-
Why do BLDC motors need electronic commutation?
No mechanical brushes. The controller must sense rotor position (Hall sensors or back-EMF) and energize correct phases.
-
What’s the purpose of the velocity loop in a servo drive?
The position loop outputs a velocity command. The velocity loop ensures the motor tracks this velocity, providing smooth motion and disturbance rejection.
-
How does reflected inertia affect motor sizing?
J_reflected = J_load / N². Higher gear ratios reduce the inertia seen by the motor, allowing faster acceleration with smaller motors.
Final Thought
“The best robot is not the one with the most powerful actuators, but the one whose actuators are precisely matched to its task, controlled with understanding, and integrated with intelligence.”
After seven weeks, you’ve built a foundation in:
- Kinematics (W2-W3): Where the robot can reach
- Dynamics (W5): What forces are required
- Sensors (W6): How the robot perceives
- Actuators (W7): How the robot moves
The remaining pieces — system integration (W8: ROS2), advanced control, and real-world deployment — build directly on this foundation. You’re ready to build complete robot systems.
Next Week (W8): ROS2 — The Robot Operating System that connects everything.