Actuators

Actuators are how computation becomes heat and motion. A DC motor does not “make torque” in the abstract; it trades electrical power through a magnetic field, with copper losses and a speed-dependent back-EMF. The datasheet curve is the whole personality: stall torque at zero speed, no-load speed at zero torque, and a continuous region you may inhabit all afternoon. Students who only ever send PWM percentages are flying without that map.

Gears and belts change the story. Reduce speed, raise torque, and you also raise the inertia the motor feels from the load — and the inertia the load feels from the motor. A high gear ratio can make a robot strong and sluggish, and it can hide a collision until something cracks. Direct drive is transparent and hungry. There is no free ratio.

Control mode is a political choice. Hobby servos hide a position loop and lie when they stall. Current control (a proxy for torque) is what force-aware arms want. On/off solenoids and pneumatics are honest about being discrete. If your algorithm assumes you can command any wrench at 1 kHz, look at the valve, the bus, and the thermal limit before you believe the API.

Other actuators exist so you do not become a motor snob. Hydraulics bring force density and leaks. Pneumatics bring compliance and slop. Series-elastic actuators put a spring where you can measure it, buying safety and losing bandwidth. Soft actuators blur the line between mechanism and muscle. Teach one electric motor deeply, then tour the zoo.

Lab that sticks: measure no-load current, stall current (briefly, with a current limit), and a mid-curve point with a known lever and scale. Fit a line. Compare to the sheet. The mismatch is wiring, temperature, or a student who trusted a 9V battery. Then read the book chapter for operating principles you should not pretend this note replaced.