Torque-Speed Characteristics of Different Motor Types

If you’ve ever spec’d a motor for a compact robot arm, a gimbal, or a prosthetic finger, you’ve stared at a torque-speed curve and felt your brain start to melt. These curves are the DNA of motion control—they tell you everything about how a motor will behave under load, at speed, and at stall. But here’s the kicker: not all torque-speed curves are created equal. In fact, the shape of that curve is the single biggest differentiator between a brushed DC motor, a stepper, a BLDC, and the unsung hero of the maker world—the micro servo motor.

This blog isn’t a textbook rehash. We’re going to rip apart the torque-speed characteristics of four major motor families, then zoom in on why micro servo motors have a unique, almost paradoxical curve that makes them perfect for precision applications. By the end, you’ll know exactly why your 9g plastic-gear servo stalls like a champ but screams like a banshee when you ask it to spin fast—and why that’s actually a feature, not a bug.

The Four Families of Torque-Speed Personalities

Before we dive into the micro servo rabbit hole, let’s map the landscape. Every rotary motor’s steady-state behavior can be plotted on a 2D plane: torque (y-axis) vs. speed (x-axis). The shape of that line—linear, drooping, flat, or wildly non-linear—tells you the motor’s “personality.” Here are the big four:

1. Brushed DC Motors: The Linear Workhorse

The classic DC motor (think hobbyist 130-size or a Mabuchi RS-540) has a linear torque-speed curve. At zero speed, you get maximum stall torque. At zero torque, you get maximum no-load speed. The curve is a straight line sloping downward from top-left to bottom-right. Mathematically, it’s T = T_stall * (1 - ω/ω_no_load).

This linearity is beautiful for control. You can predict torque at any speed with a simple equation. But the catch? Efficiency peaks at roughly half the no-load speed, and the motor’s torque drops to zero at max speed. For applications needing constant torque across a wide speed range, a brushed DC motor is a disappointment. And for micro servo motors? Well, they’re based on this linear curve, but they mutate it heavily—more on that in a minute.

2. Stepper Motors: The Flat-Torque Tyrant

Steppers are weird. Their torque-speed curve is flat at low speeds, then falls off a cliff after a certain “pull-out” frequency. Below that critical speed, a stepper can hold its full holding torque regardless of speed—that’s why they’re used in 3D printers and CNC machines. But beyond that knee point, torque plummets due to back-EMF and inductive reactance.

The problem? Steppers are open-loop position devices. They don’t have a torque-speed curve in the traditional sense—they have a torque-frequency curve. And at high speeds, they lose steps catastrophically. Micro servo motors, on the other hand, never try to be steppers. They use closed-loop feedback, so their torque-speed curve is dynamic, not static.

3. BLDC & PMSM Motors: The Efficiency Overlords

Brushless DC motors (BLDC) and permanent-magnet synchronous motors (PMSM) have a two-region curve: constant torque at low speeds (up to base speed), then constant power at high speeds (field weakening). This is the classic “EV motor” curve. It’s fantastic for traction and drones, but it requires a complex electronic speed controller (ESC) and often sinusoidal commutation.

For micro applications, BLDCs are getting smaller, but they still need external drivers. A micro servo motor, by contrast, integrates the controller, the gearbox, and the feedback sensor into a single 9-gram package. The torque-speed curve of a BLDC is smooth and efficient, but a micro servo’s curve is shaped by its gearbox and PID loop, not just its electrical constants.

4. Universal Motors: The Series-Wound Monster

Universal motors (series-wound brushed motors) have a hyperbolic torque-speed curve. Torque is inversely proportional to speed, meaning they produce massive torque at low speed and tiny torque at high speed. Think angle grinders and vacuum cleaners. They’re not precision devices. And they’re definitely not micro servo motors.

Now that we’ve set the stage, let’s dive into the star of the show: the micro servo motor. Why does its torque-speed curve look so different from a raw DC motor? And why does it matter for your next project?

Micro Servo Motors: The Closed-Loop Curve Chameleon

A micro servo motor is not a single motor. It’s a system: a small brushed DC motor (usually 7mm to 10mm in diameter), a plastic or metal gear train (reduction ratio 100:1 to 300:1), a potentiometer for position feedback, and a tiny control board that runs a PID loop. The output shaft’s torque-speed curve is not the motor’s raw curve—it’s the curve after gear reduction, plus the influence of the control loop.

Here’s the key insight: The torque-speed curve of a micro servo motor, when measured at the output shaft, is a compressed, mirrored, and non-linear version of the internal DC motor’s curve. But because the servo operates in closed-loop, the curve you measure depends on how you drive it. Let’s break this down.

The Raw Motor’s Curve Gets Gear-Multiplied

Take a typical micro servo’s coreless DC motor. It might have a stall torque of 0.5 mN·m and a no-load speed of 20,000 RPM. That’s useless for direct-drive. But with a 200:1 gearbox, the output shaft sees: - Stall torque: 0.5 mN·m × 200 = 100 mN·m (about 1 kg·cm, typical for a 9g servo) - No-load speed: 20,000 / 200 = 100 RPM (about 60 degrees per second, typical for standard servos)

So the ideal torque-speed curve at the output is still linear, but scaled. However, real gearboxes introduce friction, backlash, and compliance. That turns the pristine linear curve into a hysteretic, slightly droopy curve. At low speeds, static friction means you need a certain minimum torque to break free—this is the coulomb friction region. The curve isn’t a straight line; it’s a curve that bends downward near zero speed.

The PID Loop Rewrites the Curve

Here’s where micro servos diverge from every other motor type. When you command a micro servo to hold a position, the PID loop actively suppresses speed changes. The torque-speed curve becomes a function of the error signal. If you try to back-drive the output shaft (e.g., by pushing on the servo arm), the servo increases torque to resist. This is stiffness, not a static curve.

Graphically, if you measure torque vs. speed while the servo is trying to move at a constant angular velocity, you’ll see something bizarre: - At zero commanded speed (holding position), the servo can produce its full stall torque in both directions, but only if the error is large enough. The curve looks like a vertical line at speed = 0. - At moderate commanded speeds (say, 50% of max speed), the curve is steep but not linear. The PID’s proportional term dominates, so torque drops off faster than the raw motor’s linear curve. - At near-max commanded speed, the integral term winds up, and the servo may overshoot—the curve can actually show negative torque for a brief moment as it corrects.

This is a speed-dependent, non-linear, closed-loop torque-speed characteristic. No other motor type behaves this way. A stepper’s flat curve is open-loop. A BLDC’s curve is open-loop (with ESC). A micro servo’s curve is alive—it changes shape based on the gain settings, the load inertia, and the commanded trajectory.

Why the Curve Makes Micro Servos “Torque-Biased”

Let’s talk numbers. A typical SG90 micro servo has a stall torque of 1.8 kg·cm at 4.8V. Its no-load speed is 0.1 sec/60° (about 100 RPM). Now, plot that on a torque-speed graph. The stall torque is high, but the no-load speed is pathetically low compared to a raw DC motor or a BLDC. The curve is steep—meaning torque drops off very quickly as speed increases.

Why is this a good thing? Because micro servos are designed for position control, not velocity control. You rarely need a servo to spin continuously at 100 RPM. You need it to hold a load at a specific angle, or to move in small increments with high precision. A steep torque-speed curve means: - High holding stiffness: Small perturbations in load don’t cause large speed deviations. - Predictable settling: The servo reaches its target angle with minimal overshoot because the torque available at high speed is low, naturally damping the motion. - High peak torque at stall: For short bursts (e.g., holding a robot arm against gravity), the servo can deliver its maximum torque even at zero speed.

Compare that to a BLDC with a flat constant-torque region. If you tried to use a BLDC with a gearbox for position control, you’d need a very aggressive PID loop to prevent overshoot, because the motor can produce high torque even at high speed. The micro servo’s inherent torque-speed droop acts as a mechanical damper.

The “Dual Curve” Phenomenon: Static vs. Dynamic Torque

One of the most misunderstood aspects of micro servo torque-speed characteristics is the difference between static (stall) torque and dynamic (running) torque. For a raw DC motor, stall torque is a single point on the curve. But for a micro servo, the effective stall torque depends on the PWM pulse width and the current limit.

The PWM Voltage Bucket

Micro servos are typically driven by a 50Hz PWM signal with a pulse width between 1ms and 2ms. The servo’s internal controller translates that pulse width into a target position, then applies full supply voltage to the motor until the error is zero. This is a bang-bang with PID control, not a smooth voltage ramp.

So, the torque-speed curve you’d measure at the output shaft is actually a family of curves, one for each pulse width. At 1.5ms (center position), the servo will hold torque in either direction. At 1.0ms (full left), the servo will only produce torque in one direction—and if you back-drive it, it will fight you until it hits the mechanical stop.

This means you can’t define a single torque-speed curve for a micro servo. Instead, you define a torque-speed envelope. Within that envelope, the servo can produce any combination of torque and speed, but the shape of the envelope is constrained by: - The motor’s raw stall torque and no-load speed (after gear reduction). - The battery voltage (higher voltage = higher stall torque and no-load speed). - The current limit on the controller (often 1A to 2A for micro servos). - The gearbox efficiency (typically 70-80% for plastic gears, 85-90% for metal gears).

A Practical Example: The 9g Servo vs. A Micro BLDC

Let’s compare a 9g micro servo (e.g., SG90) with a 10mm BLDC motor (e.g., a drone motor) that has been gear-reduced to the same speed range. The BLDC might have a stall torque of 2 kg·cm and a no-load speed of 150 RPM. Its torque-speed curve is relatively flat up to 100 RPM, then falls off.

Now, command both to move from 0° to 90° in 0.2 seconds. The micro servo will: - Accelerate quickly to its max speed (~100 RPM), but as it approaches the target, the PID reduces torque, and the speed drops non-linearly. The result is a smooth, critically-damped response. - The BLDC will accelerate to its max speed, but because its torque remains high at speed, it will overshoot the target, then oscillate. You’ll need a sophisticated controller to damp that.

The micro servo’s steep torque-speed curve is its secret weapon. It’s not about raw power—it’s about controllability.

The Gearbox’s Hidden Influence on the Curve

You can’t talk about micro servo torque-speed characteristics without dissecting the gearbox. Most micro servos use planetary gearboxes (2-stage or 3-stage) or spur gear trains. The gear ratio directly scales the torque-speed curve, but it also adds three nasty artifacts:

1. Backlash Dead Zone

Backlash is the play between gear teeth. On the torque-speed curve, backlash creates a dead zone near zero torque. When the servo reverses direction, the output shaft doesn’t move until the gears re-engage. This shows up as a flat spot on the curve at low torque values—the servo produces zero output torque for a small angular displacement. This is why micro servos feel “sloppy” when they’re cheap. High-end metal gearboxes reduce backlash to <1°, which tightens the dead zone.

2. Friction Hysteresis

The gearbox adds static friction. When you measure torque-speed by slowly rotating the output shaft, you’ll see different curves for increasing vs. decreasing speed. This hysteresis loop is caused by Coulomb friction in the gears and bearings. The width of the loop is proportional to the friction torque. For a micro servo, this hysteresis can be as high as 10-15% of stall torque, meaning the servo’s effective torque at very low speeds is unpredictable.

3. Elastic Compliance

Plastic gears flex under load. This compliance acts like a spring between the motor and the output shaft. On the torque-speed curve, this manifests as a phase lag—the output speed lags behind the commanded speed at high torque. This is why a 9g servo under heavy load will appear to “stutter” at low speeds; the gear train is winding up and releasing.

Temperature’s Effect on the Curve: The Invisible Killer

Here’s a curveball: a micro servo’s torque-speed characteristic is not static over time. As the motor heats up, the resistance of the windings increases. This reduces the stall torque (since T_stall = Kt * V / R) and increases the no-load speed (since ω_no_load = V / Ke). The entire curve rotates clockwise—you get less torque at every speed.

But that’s not all. The gearbox’s lubricant thins out as it warms up, reducing friction torque. So the hysteresis loop narrows. The net effect is that a cold micro servo feels stiff and slow, while a hot one feels loose and fast. If you’re designing a system that runs for hours, you need to account for this thermal drift in your torque-speed model.

For micro servo motors, this is especially critical because they’re often used in high-duty-cycle applications (e.g., animatronics, camera gimbals). A servo that holds a load for 10 minutes can lose 20% of its stall torque due to heating. The torque-speed curve you measured on the bench at 25°C is a lie at 70°C.

Comparing Micro Servo Curves to Other “Small” Motors

Let’s put the micro servo head-to-head against two other popular small motors: a coreless DC motor (like the ones in phone vibrators) and a micro stepper (like the 28BYJ-48).

Micro Coreless DC Motor (e.g., N20 with gearbox)

These are essentially micro servo motors without the feedback and controller. Their torque-speed curve is linear (after gear reduction), but there’s no closed-loop shaping. You can drive them with a simple H-bridge, but you get zero holding torque. If you stall them, they overheat. The curve is predictable but dead—no stiffness, no damping.

Micro Stepper (e.g., 28BYJ-48)

This stepper has a flat torque curve up to ~10 RPM, then it nosedives. Its holding torque is excellent (you can just leave it powered), but its dynamic torque at moderate speeds is poor. Worse, it has a resonance region where torque drops to near zero. The micro servo’s torque-speed curve, by contrast, has no resonance because the PID loop actively damps oscillations.

The Verdict

The micro servo motor’s torque-speed characteristic is a hybrid: it borrows the linearity of a DC motor, the holding stiffness of a stepper, and the closed-loop precision of a BLDC, but it sacrifices top speed and continuous torque. For applications that require high torque at low speed, precise position holding, and compact size, the micro servo is unmatched. Its curve is steep, non-linear, and temperature-dependent, but that’s exactly what makes it so good at what it does.

How to Read a Micro Servo Datasheet (Without Being Fooled)

Every micro servo datasheet lists two numbers: stall torque and speed. But these are no-load speed and stall torque—the two endpoints of the linear curve. The actual curve in between is rarely shown. Here’s how to decode the marketing:

• Stall torque at 6V vs. 4.8V: Most servos are rated at 4.8V, but running at 6V increases stall torque by ~25% and speed by ~20%. The curve shifts outward.
• Speed is always no-load: The “0.1 sec/60°” spec is with zero load. Under 50% load, that time doubles. The torque-speed curve is not a straight line to the stall point; it’s concave down due to friction.
• Stall current vs. running current: The stall current can be 5-10x the running current. This means the torque-speed curve is not only steep, but the power curve peaks at about 50% of no-load speed, not at stall. For micro servos, operating above 50% speed is inefficient.

If you want to measure the real curve, use a torque sensor and a rotary encoder, and sweep the commanded position slowly while recording the torque. You’ll get a closed-loop hysteresis loop, not a line. That loop’s width is the servo’s “personality.”

Why Micro Servo Torque-Speed Curves Matter in Real Applications

Let’s ground this in three real-world examples:

1. Robot Finger Actuator

A micro servo in a robot finger needs to close quickly (high speed) and then hold a grip (high stall torque). The steep torque-speed curve means the servo can snap to a target angle fast, but as it approaches the target, the torque drops, preventing the finger from slamming into the object. The gearbox’s compliance acts as a shock absorber. A BLDC would crush the object.

2. Camera Gimbal (Pitch Axis)

A gimbal motor needs to produce constant torque at zero speed to hold the camera level. A micro servo’s closed-loop curve is perfect here—it can hold torque indefinitely at speed = 0 without overheating (as long as the stall current is within limits). The hysteresis loop means the servo won’t oscillate around the level point; it will settle into a stable dead zone.

3. Continuous Rotation Servo (Modified)

Some users remove the position feedback to make a continuous-rotation servo. The torque-speed curve becomes exactly the raw DC motor’s linear curve (after gear reduction), but without the PID loop, it loses all its damping. This is why modified servos are terrible for precision—they’re just gear-motors with a built-in ESC.

The Bottom Line (But Not a Conclusion)

The torque-speed characteristics of different motor types are not just academic curves—they’re the fingerprints of each motor’s soul. Brushed DC motors are linear and predictable. Steppers are flat and stubborn. BLDCs are efficient and flat until they’re not. But micro servo motors? They’re chameleons. Their curves are steep, non-linear, temperature-sensitive, and heavily influenced by the gearbox and PID loop. That steepness is a feature: it gives them inherent damping, high holding stiffness, and precise positional control at the expense of top speed and continuous power.

So the next time you pick up a 9g servo and feel its resistance when you turn the horn by hand, remember: you’re feeling the combined effect of a gear-reduced linear curve, a PID loop fighting you, and the friction of plastic gears. That’s not a flaw—that’s the torque-speed characteristic of a motor type that chose control over speed, and that’s exactly why it’s still the go-to for every hobbyist, roboticist, and animatronics engineer on the planet.