Cracking Actuator Cost Challenge

•

Actuators eat up 30-50% of a humanoid robot's BOM cost. Mass adoption? Not at these prices. This is one of the most persistent challenges in robotics, but @IMSystemsNL may have cracked the code with a novel drive. To understand it, we need to dive into the world of drives & transmission, with all their demands and trade-offs. Shoutout to @MarwaEldiwiny and @GoingBallistic5 for the original video.

The actuator dilemma that's plagued industrial robotics for 50 years: Electric motors are most efficient at high speeds (5,000-20,000 RPM). But robot joints need slow, controlled movement (~60 RPM max). This forces massive gear reductions (50:1 to 320:1) - and that's where everything breaks down.

Why massive gear reductions suck: - Backlash multiplies: Every tooth adds play - multiply by 50x-320x reduction and your robot can't draw a straight line - Efficiency drops: Each stage loses 2-5% efficiency - Weight explodes: More stages = more mass to accelerate - Back-driveability vanishes: High ratios make manual joint movement impossible - a safety nightmare The bigger the reduction, the worse these problems get.

So, how do engineers solve this? Three dominant approaches: • Planetary gears: Stack multiple stages for high ratios • Harmonic drives: Achieve huge ratios in single stage via flex • Cycloidal drives: Use eccentric motion for compact reduction (Excludes belts, tendon, worm gears, ball screws, fluid) Each solves some problems while creating others.

Planetary gears work like a solar system - central "sun" gear surrounded by "planet" gears inside a ring. Great efficiency (95-98%) but the flaw: tiny gaps between teeth multiply by 50x reduction and your robot can't draw a straight line.

Harmonic drives flex a thin metal cup with an oval cam to create massive reductions in one step. Zero slop, but they cost $1,000-$3,000 each and only 3-4 companies worldwide can make them properly. Supply chain bottleneck.

Cycloidal drives use an off-centre disc that "wobbles" against pins for reduction. Super rigid and strong, but complex to manufacture. You need aerospace-level precision machining.

This is an even bigger problem for humanoids that have even tougher constraints: 1. Safety (must not injure humans during interaction) 2. Cost (sub-$20,000 robot requires ~$500 actuators) 3. Performance (zero backlash, high bandwidth, force transparency, low weight) Current transmissions force trade-offs. Humanoids need all three.

This is why the industry is chasing direct drive - motors directly powering joints with no transmission at all. Direct drive would solve everything: zero backlash, perfect force transparency, infinite stiffness, no gear complexity. It's the holy grail. But no one has cracked it due to motor size and cost constraints. The compromise? Quasi-direct drive (QDD) with low gear ratios (6:1 to 20:1). Good but not perfect.

Enter the Archimedes Drive from IMSystems—a departure from gear-based thinking. Instead of metal teeth, they use controlled friction between rolling cylinders under compression. Think: train wheels on tracks, but in a planetary configuration.

The physics are elegant: • Hollow rolling cylinders replace gears • Pre-compression creates spring-loaded contact • Force transfers through traction, not gear mesh • Variable geometry enables ratios from 6:1 to 100:1+ Think: infinite gear teeth at the microscopic level.

Three breakthrough characteristics: • Zero backlash from continuous rolling contact • Inherent slip protection - overtorque causes slipping, not breaking • Perfect back-driveability for human-safe interaction You can hit it with a hammer and it just slips safely at preset limits.

Manufacturing advantages: Traditional gears need specialised grinding equipment ($70K-$100K per tool) with limited suppliers. The Archimedes Drive uses bearing manufacturing - a mature, scalable industry. Early estimates suggest 50-80% cost reduction vs harmonic drives.

The counterintuitive insight: what made this drive "unsuitable" for industrial robots makes it ideal for humanoids. @IMSystemsNL spent 10 years being "too early" - industrial robots saw slip as a liability. Now humanoid companies specifically request slip protection for safety. Yesterday's weakness became today's advantage.

Prediction: Within 5 years, two distinct ecosystems: • Industrial: High-ratio drives optimised for precision • Humanoid: Low-ratio friction drives optimised for safety and cost The Archimedes Drive represents a step in this direction.

The bigger picture: We're at an inflection point similar to the transition from mechanical to electronic systems. Is the next chapter isn't about making better gears or about eliminating gears? Here's the link to the full video: youtube.com/watch?v=XqOumFdvZps&t=573s&ab_channel=MarwaElDiwi

Want access to the latest robotics jobs? 🌐 Our website updates with new roles every Friday 📧 Newsletter subscribers get early access on Wednesday Stay ahead of the curve → robot-iq.org/