Latest / The Robotics Podcast with Fexingo: Autonomous Systems, Industrial Robots, and Hardware / Why Solar Panels Still Dwarf the World's Most Efficient Robots
Transcript
- Lucas: So I'm standing in front of a typical six-axis industrial robot arm right now — the kind you'd see welding car frames or palletizing boxes — and I'm looking at its nameplate power rating. It's about seven kilowatts. Luna: Seven kilowatts — that's like running three household ovens at once. Lucas: Exactly. And this isn't some giant mining excavator. This is a mid-size arm that lifts maybe ten kilograms. And it's running eight, sometimes sixteen hours a day. So the energy cost isn't trivial — especially when you multiply across thousands of units in a factory. Luna: I remember reading that a single Tesla Gigafactory can draw as much power as a small city. And a lot of that goes to robots, not just the battery lines. Lucas: Right. Which brings us to the topic today — robot energy efficiency. Or more bluntly: why most robots are still energy hogs, and how a few companies are starting to fix that. Luna: And why this matters beyond the electricity bill. Lucas: Exactly. Because energy consumption feeds into heat management, battery life for mobile robots, and even the structural design of the robot itself. If you can cut power draw by even twenty percent, you might not need that heavy cooling system, and suddenly the robot can be lighter, cheaper, more agile. Luna: So what's the baseline? How inefficient are we talking? Lucas: Well, a human worker doing equivalent light assembly burns about one hundred watts of metabolic power. That's roughly the same as a bright incandescent light bulb. This robot arm next to me draws seven thousand watts to do a similar physical task — and it's often slower and less dexterous. Luna: That's a factor of seventy times more energy per task. That is staggering. Lucas: And humans don't need a twenty-kilowatt cooling system to keep their joints from overheating. The robot does. Most of that seven kilowatts isn't doing useful mechanical work — it's lost as heat in the motors, gearboxes, and power electronics. Luna: So the efficiency problem is partly a physics problem — electric motors are around ninety percent efficient, but gearboxes and friction eat into that. Lucas: Right. And the rest is system design. Most industrial robots use heavy, rigid arms with large servomotors because they have to move fast and precisely. But that means every move accelerates and decelerates a lot of mass — and that takes energy. Luna: So the lighter the arm, the less energy needed. But you can't just make it lighter if it needs to be stiff enough to position a heavy payload accurately. Lucas: There's a trade-off. But some engineers are challenging the assumption that you need a rigid arm at all. A company called Haddington Dynamics — based in Las Vegas — has developed a robot arm that uses cables instead of gears. Think of it like a marionette, but with high-tension cables and lightweight carbon-fiber links. Luna: I've seen those — they look almost skeletal. Does it actually work in a factory? Lucas: It does. Haddington's arm draws about two kilowatts peak instead of seven, and it's accurate to within a tenth of a millimeter. They've deployed it in electronics assembly and even for surgical robot research. The key is that the motors stay in the base — they don't move with the arm — so you're not accelerating the motor's own mass with every joint movement. Luna: That's clever. So it's like moving the engine out of the car and into the garage, and just having cables pull the wheels. Lucas: Exactly that analogy. And because the arm itself is lighter, you can use smaller motors, which reduces cost and power draw. The company claims a seventy percent reduction in energy consumption compared to a comparable conventional arm. Luna: Seventy percent — that's not incremental, that's a step change. Lucas: If today was actually useful to you, the way these episodes stay ad-free is listener support. You can find it at buy me a coffee dot com slash fexingo. Luna: Yeah, it's a small way to keep this kind of deep dive going. Lucas: So back to cables. The trade-off is that cable-driven arms have lower stiffness than rigid gear arms. If you need to do heavy milling or apply hundreds of newtons of force, cables won't cut it. But for pick and place, assembly, or even light sanding, it works beautifully. Luna: So it's not a universal replacement. But it's a reminder that the dominant robot design — the big cast-iron six-axis arm — was optimized for the 1980s automotive industry, not for modern, flexible manufacturing. Lucas: Exactly. And energy is becoming a bigger factor as companies face pressure to reduce carbon footprints and cut operating costs. A typical factory pays for electricity every month, and those kilowatt-hours add up. Luna: I've seen numbers that energy can be ten to fifteen percent of the total cost of ownership for an industrial robot over its lifetime. Lucas: And that's just the direct cost. There's also the indirect — heat generation, cooling requirements, and even the size of the power distribution system in the factory. If every robot draws half the power, you might need fewer transformers and smaller cables. Luna: So the efficiency improvements cascade through the whole factory design. Lucas: Right. And some large manufacturers are starting to require energy-efficiency data when they evaluate robot suppliers. BMW, for example, has an internal standard for energy per part produced. Luna: That's a powerful incentive. If you're a robot maker and your arm uses thirty percent less power, you can win a contract on that alone. Lucas: Which brings us to another approach: software optimization. A company called Rockwell Automation has shown that by optimizing the motion path — the trajectory the robot takes between points — you can reduce energy consumption by up to twenty percent without any hardware changes. It's like driving smoothly instead of accelerating hard and braking. Luna: So the same robot, just smarter motion planning. Lucas: Yes. And it's not just about speed. You can also time the movements to avoid simultaneous high-torque demands on multiple joints. That spreads the peak power load and reduces the overall energy per cycle. Luna: That sounds like something machine learning could optimize automatically. Lucas: It already is. There's research from MIT and some startups using reinforcement learning to find energy-optimal trajectories. The robot runs a few thousand simulated cycles, learns the most efficient path, and then executes that in production. Luna: And it updates as the robot's motors wear and friction changes. Lucas: Exactly. So we're moving from a world where robot efficiency was an afterthought — 'just plug it in' — to one where it's a design criterion from the start. Luna: I wonder if we'll see energy-efficiency labels on robots, like we do on refrigerators. Lucas: That's not far-fetched. The European Union already has a directive on energy-using products that covers industrial equipment. I could see an 'A++' rating for robot arms within five years. Luna: So the question is: will hardware innovation or software optimization win the race? Lucas: Probably both. Cable-driven arms are great but niche. Motion optimization is universal but limited to maybe thirty percent savings. The real step change might come from new actuator technologies — like direct-drive motors that eliminate gearboxes entirely, or even piezoelectric motors that are tiny and efficient. Luna: But those are still expensive and not very powerful. Lucas: True. For now, the biggest lever is simply awareness. Most factory managers don't track energy consumption per robot. They just see the total plant electricity bill. If they started metering individual arms, they'd find the worst offenders and either replace them or reprogram them. Luna: And that's a low-hanging fruit. Lucas: Exactly. So if you're in manufacturing, my advice for this week: look at your robot nameplates. If you see anything above five kilowatts for a small arm, ask your team if the trajectory is optimized. You might save more than you expect. Luna: And if you're designing a robot from scratch, think about where the motors live. Lucas: Right. The cable-driven approach or even a simple belt drive could make a difference. The era of the energy-hungry robot is not over, but it's definitely on notice.