Back to blog overview

Thermal management and zero-power hold

Packing over a dozen active degrees of freedom into a palm and fingers the size of a human hand creates a thermal management challenge. Every joint requires an actuator, but inside a compact finger or wrist, there is virtually no room for cooling fans, liquid cooling loops, or heat sinks.

Energy is lost during motion, when the actuator moves. And it is lost when the robot is stationary, because of static holding power consumed when maintaining a posture.

At a small enough size, an EM motor becomes a heater

EM motors operate on the Lorentz force, where an electric current I flowing through a conductor of length L within a magnetic flux density B generates a mechanical force:
$$\vec{F} = I (\vec{L} \times \vec{B})$$

On macroscopic scales, electric motors are remarkably efficient—often exceeding 90% conversion efficiency. However, when motor dimensions shrink down to the millimeter scale, scaling laws become brutally unfavorable.

As the size of an inductive coil shrinks, the volume available for copper wire decreases rapidly, requiring much thinner wire gauges to maintain the required turn count. The electrical resistance R of these miniature coils scales inversely with wire cross-sectional area. Resistive losses dominate at sub-gram mass scales.

According to motor characterization metrics established by Harrington and Kroninger [1], at a mass of 2 grams, maximum practical motor efficiency drops to 50%. At a mass of 0.5 grams, motor efficiency plummets to 0.5%. At these sizes, an EM motor essentially becomes an electrical heater.

The high cost of doing nothing

While active motion generates significant heat, robotic hands spend a vast majority of their operational life holding static postures—pinching a tool, holding an object against gravity, or maintaining a firm grip. 

In an EM motor, maintaining a static holding torque requires continuous electrical current I  through the stator windings to produce counter-torque against the load:
$$P_{\text{static}} = I^2 R > 0$$

Because no mechanical work is performed while stationary, all of the electrical power drawn during a static hold is dumped directly into the hand as waste heat. To keep a 15-joint robotic hand gripping an object, every single motor continuously consumes battery power and raises internal temperatures, even though the hand is entirely motionless.

Electrostatic vs. electromagnetic forces: a fundamental shift

To solve both the motion cliff and the stationary heating problem, roboticists have long eyed electrostatic actuation  as an alternative. The fundamental physics governing electromagnetic and electrostatic motors stem from two entirely different forces:

Dimension

Electromagnetic motor

Electrostatic motor

Principle

electromagnetic

Lorentz Force: Electric current I passing through magnetic flux B.
$\vec{F} = I (\vec{L} \times \vec{B})$

electrostatic

Coulomb's Law: Electric field E acting on static charges q across capacitance C.
$\vec{F} = q\vec{E} = \frac{1}{2} \frac{dC}{dx} V^2$

Static holding mechanism

electromagnetic

Requires continuous current flow I through resistive coils

electrostatic

Pure capacitive charge storage (Q = CV) on dielectric surfaces

Static-hold power draw

electromagnetic

High (P = I²R) heat generation

electrostatic

Zero; pure electrostatic field retention

Miniaturization

electromagnetic

Unfavorable; coil resistance R dominates at microscale

electrostatic

Highly favorable; capacitive surface-to-volume ratio increases

Because electrostatic motors rely on static electric charges stored across capacitive surfaces rather than continuous current flowing through resistive wires, maintaining a static position requires zero continuous current flow. Once the capacitive plates are charged to a target voltage, the electrostatic attraction force holds the position indefinitely without drawing active power or generating Joule heat.

Why conventional electrostatic motors fall short

Despite their theoretical appeal, conventional electrostatic motors—first explored by Benjamin Franklin and Andrew Gordon in the late 1740s, and later revived in MEMS comb-drives during the late 20th century—have historically failed to displace electric motors. Conventional electrostatic approaches suffer from severe physical limitations:

Low force and torque density in air: The maximum electric field strength in air is capped by dielectric breakdown (3 MV/m). Because electrostatic pressure is proportional to the dielectric constant, conventional air-gap electrostatic motors produce extremely low forces per unit area, and very low torque density.

High operating voltages: Achieving meaningful displacement or force across air gaps usually requires voltages in the hundreds or thousands of volts, requiring bulky high-voltage boosting electronics.

Inability to scale volumetrically: Traditional MEMS comb-drives are fabricated as thin, planar silicon structures. Because they cannot be easily stacked into thick 3D structures, their total force output remains trapped at the micro-Newton level.

Stiction, friction, and wear: Solid-on-solid contact between moving microscopic parts leads to severe stiction and mechanical abrasion, destroying operational reliability.

Pull-In instability : Flexible dielectric actuators often suffer from pull-in breakdown, where electrostatic forces collapse the thin air/dielectric gap, causing catastrophic short circuits.

Microhydraulic actuators: All of the benefits, none of the limitations

Microhydraulics transforms electrostatic actuation by replacing rigid air gaps or dry solid contacts with microscopic liquid droplets surrounded by insulating oil on polyimide thin-films. Applying a modest voltage alters the surface tension of the droplets via electrowetting on dielectric (EWOD), deforming the droplets and pulling adjacent layers with immense force.

Here is how microhydraulic actuators overcome the historical limits of both electromagnetic and electrostatic motors.

Zero-power static hold

In Sangtera actuators, the electrical charge is stored across a high-quality solid/fluid capacitive interface. 98% of the applied voltage is dropped across the solid electrowetting dielectric [2]. Once the desired posture or grip force is reached, the drive electrodes hold their capacitive charge Q = CV. Because the dielectric exhibits virtually zero leakage, the actuator maintains full holding force at zero DC power draw, eliminating static thermal dissipation in robotic fingers.

1/D² Quadratic Scaling

Unlike electromagnetic motors that degrade at small scales, Sangtera actuators exhibit quadratic power and torque density scaling as internal dimensions shrink. The smaller the droplet pitch—the distance between two adjacent droplets—the more powerful the motor.

Both the driving capillary pressure  and the maximum cycle frequency  scale inversely with droplet pitch D. Consequently, the power density is proportional to 1/D²; halving the droplet pitch quadruples power density.

At a 40-micron droplet pitch, Sangtera achieves a  stall torque density of 78 Nm/kg at just  50V. This is more than 100 times higher than the torque density of sub-centimeter electromagnetic stepper motors [3], such as the Precision Microdrives 303-102 at 0.68 Nm/kg².

Actuator technology

Sub-centimeter EM stepper motor

MEMS electrostatic motor

40 µm pitch microhydraulic motor [4]

Principle

sub-cm EM stepper

Lorentz Force

MEMS electrostatic

Coulomb's Law

40-µm pitch microhydraulic

Electrowetting capillary pressure

Operating voltage

sub-cm EM stepper

3-12 V

MEMS electrostatic

150V to several kilovolts

40-µm pitch microhydraulic

50V

Stall torque density

sub-cm EM stepper

0.68 Nm/kg

MEMS electrostatic

< 0.1 Nm/kg

40-µm pitch microhydraulic

78 Nm/kg

Static-hold power draw

sub-cm EM stepper

High P = I²R

MEMS electrostatic

Zero I = 0

40-µm pitch microhydraulic

Zero I = 0

3D Stackable Thin-Film Architecture & Liquid Interconnects

To produce macro-level mechanical forces, Sangtera integrates electrode and droplet arrays onto individual polyimide wafers that are 10 μm thick. To power multiple sliding layers without wires, Sangtera uses an ingenious  liquid interconnect network: built-in saltwater rails and vertical fluidic vias carry four-phase AC electrical signals straight through the moving stack, enabling true 3D volumetric actuation.

Elimination of Stiction and Pull-In Instability

Because every moving polyimide layer is separated by a continuous liquid film of droplets and oil, solid surfaces never rub against each other, eliminating stiction and mechanical wear. Furthermore, because the solid polyimide dielectric is rigid, the electric field remains stable during operation, preventing the pull-in breakdown that plagues compliant electrostatic actuators.

The Future of Dexterous Microrobotics

By addressing both sides of the thermal management equation, microhydraulic electrowetting actuators redefine what is possible in compact robotics:

Elimination of Motion Overheating: Operating at over 60%  energy conversion efficiency, microhydraulic actuators deliver high mechanical power without heating.

Zero-power static hold: Gripping objects for extended periods draws zero power, preserving battery life and keeping joint temperatures at ambient levels.

High torque without gear: High torque density allows robotic fingers to exert real strength without heavy, space-consuming gear trains.

[1] In 2013, Aaron M. Harrington and Christopher Kroninger established that standard manufacturer metrics like the KV rating are insufficient for predicting actual operational performance under load. Instead, they identified key electromechanical parameters that must be measured simultaneously using a power analyzer and a dynamometer to properly characterize small DC brushed and brushless motors.

[2] Jakub Kedzierski and Hero Chea, "Multilayer microhydraulic actuators with speed and force configurations. Microsystems & Nanoengineering 7, 22 (2021). DOI: 10.1038/s41378-021-00240-7

[3] Jakub Kedzierski and Eric Holihan, "Linear and rotational microhydraulic actuators driven by electrowetting." Science Robotics, Vol 3, Issue 22 (2018), DOI: 10.1126/scirobotics.aat5643

[4] We only cite numbers that are in the public domain. Specifications have since evolved.

[5] Ibid.

Read more

100x increase in torque density

Calculations that back up our performance claims.

Read the post

Tradeoffs in building dexterous hands

Current approaches, their limitations, and how Sangtera’s actuators overcome them.

Read the post

The QDD paradigm

An overview of quasi-direct drives. How they dominate humanoid legs, and Sangtera brings them to fingers.

Read the post