# How Chip-Scale Galvos Are Shrinking Laser Scanners

An episode of [The Hardware Podcast with Fexingo: Chips, Devices, and Electronics Engineering Conversations](https://invisiblenetwork.ai/podcast/1896799982), 21 Aug 2026, 9:31.

Audio: https://audio.fexingo.com/business/the-hardware-podcast/episode-0162.mp3

In this episode of The Hardware Podcast, Lucas and Luna explore how chip-scale galvanometer mirrors are making laser scanning systems dramatically smaller, faster, and cheaper. They trace the technology from traditional bulky galvos used in industrial engraving and lidar to the new MEMS-based designs now appearing in AR glasses, robotic vision, and medical imaging. Lucas explains the physics of electrostatic comb drives and how they achieve high-speed beam steering with minimal power, while Luna highlights a specific example: a MEMS galvo from a recent research prototype that fits on a fingertip yet scans at kilohertz rates. They discuss the trade-offs between bandwidth, mirror size, and optical quality, and why the thermal and mechanical constraints are different from conventional galvos. The conversation also touches on packaging challenges, yield issues, and what needs to happen for these chips to move from labs to mass-market devices. A short, natural support segment near the end ties the discussion of shrinking technology to keeping the show ad-free. #ChipScaleGalvos #MEMS #LaserScanning #Lidar #AugmentedReality #Robotics #MedicalImaging #ElectrostaticActuators #BeamSteering #Microfabrication #HardwareEngineering #MicroOptics #FexingoBusiness #BusinessPodcast #TechnologyPodcast #Engineering #Semiconductor #DeepTech Keep every episode free: buymeacoffee.com/fexingo

## Transcript

Lucas: So we've spent a lot of episodes on chip-scale sensors and clocks and spectrometers, but today I want to talk about something that acts. We're talking about chip-scale galvanometer mirrors, the tiny moving mirrors that steer laser beams without any bulky external optics.
Luna: Galvos. I always think of those big industrial boxes in laser engraving machines, like the size of a toaster. So how small are we talking exactly?
Lucas: Smaller than a fingernail. The whole mirror and actuator fits on a chip maybe five millimeters on a side. And the mirror itself is often less than a millimeter across. But it can move fast enough to draw a high-resolution pattern or scan a 3D scene.
Lucas: Think of a traditional galvo: a motor, a mirror, a position sensor, all machined metal. They're precise, but they're heavy and they draw watts. A MEMS galvo replaces that with a silicon mirror suspended on tiny flexures, driven by electrostatic forces. No motor, no coil.
Luna: And electrostatic means high voltage, low current, right? I remember reading about comb drives in MEMS accelerometers.
Lucas: Exactly. Interdigitated comb fingers, like a tiny comb, that pull on the mirror when you apply a voltage. The voltage can be tens of volts, but the current is minuscule, so the total power is often in the milliwatt range. That's a game changer for battery-powered devices.
Luna: So where are we actually seeing these in products? Not just in a lab?
Lucas: The most visible place is in lidar for autonomous vehicles and robots. A solid-state lidar can use a MEMS mirror to steer a laser across the field of view instead of spinning a whole assembly. There are also AR glasses, like the ones from a few startups, that use tiny mirrors to project images directly onto your retina.
Luna: Retina projection, that's wild. And I've seen some interest in medical imaging, like endoscopic OCT. That must really need a tiny scanner.
Lucas: Absolutely. For OCT, you want a probe small enough to go through a catheter, and a MEMS mirror can fit in a package that's about a millimeter in diameter. It can scan at kilohertz rates, which is fast enough to capture real-time cross-sections of tissue.
Luna: So what's the catch? Why isn't every laser scanner a MEMS chip by now?
Lucas: The big challenge is the trade-off between mirror size and bandwidth. A larger mirror reflects more light and has a better optical quality, but it's heavier and slower. You have to find the sweet spot for each application. And then there's the issue of tilting angle. Some applications need a wide optical scan, like more than 30 degrees, which is hard for a small mirror.
Lucas: There's also the packaging problem. You have to seal the mirror in a hermetic package with a window, and that window needs to be anti-reflection coated for the exact wavelength you're using. It's not trivial, and it adds cost.
Luna: And what about reliability? Moving parts in a MEMS device, even if they're tiny, must suffer from fatigue or stiction.
Lucas: Silicon actually has excellent fatigue resistance, much better than steel, so the flexures can survive billions of cycles. Stiction is a concern, but that's usually a manufacturing problem, not a field failure. You design the springs stiff enough and you use anti-stiction coatings.
Lucas: The real issue is yield. When you're making a mirror that has to be perfectly flat to within a fraction of a wavelength of light, any residual stress in the thin film can warp it. That's why you see a lot of mirrors made from single-crystal silicon, which has no internal stress.
Luna: So the material choice is critical. I've heard of some MEMS mirrors using aluminum or gold for the reflective coating, but the structural layer is usually silicon.
Lucas: Right. And you have to balance reflectivity with the thermal expansion. If the mirror absorbs even a tiny fraction of the laser power, it can heat up and deform. For high-power lasers, that's a problem. But for low-power applications like lidar or AR, it's manageable.
Luna: Let's talk about a specific example. I recall a research prototype from a university a couple of years ago that achieved a scan rate of over 10 kilohertz with a mirror about a millimeter wide. How does that compared to commercial parts?
Lucas: Ten kilohertz is impressive. Most commercial MEMS mirrors are in the range of one to five kilohertz, and they often sacrifice angle to get that speed. That research prototype probably used a resonant drive, where you oscillate the mirror at its mechanical resonance frequency, which allows large angles but only at a fixed frequency and sinusoidal waveform.
Lucas: For raster scanning, you often use two mirrors, one fast and one slow, or you use a single mirror with a Lissajous pattern. The control electronics have to be carefully designed to stabilize the amplitude and compensate for temperature drift.
Luna: So it's not just the mirror, it's the whole control loop. And that's where the real engineering challenges lie.
Lucas: Exactly. And that's why we're seeing a lot of startups emerge that combine MEMS design with the drive electronics and the software to make the whole system work. The mirror is just the heart, but you need the brain and the nervous system around it.
Luna: I'm thinking about the cost, too. A traditional galvo assembly can cost hundreds of dollars. What's the price point for a MEMS mirror in volume?
Lucas: In high volume, the MEMS chip itself might be five to ten dollars, and the whole module, including the packaging and the driver, maybe twenty to thirty. That's a huge reduction, and it's what's enabling lidar to drop from thousands of dollars to a few hundred, which is why we're seeing it in more consumer robotics.
Lucas: But there's still a gap between the promise and the reality. Yield issues and the need for custom calibration can eat into those savings. But the trend is clear: everything is shrinking, and the cost curve is falling.
Luna: It feels like we're at that inflection point where MEMS mirrors become a commodity, kind of like accelerometers did a decade ago.
Lucas: That's a great comparison. And when that happens, we'll see them in everything from cheap 3D scanners to smart home devices. It's one of those quiet technologies that will just show up in the background.
Luna: You know, this episode reminds me of how much I appreciate the depth we can get into on this show. And speaking of supporting that depth, I know we have a way for listeners to keep this going.
Lucas: Yeah, if today's conversation gave you something useful, and you want to help us keep the show independent and ad-free, you can support us at buy me a coffee dot com slash fexingo. It's a simple way to chip in, and it really helps us invest in research and bring you more episodes like this one.
Luna: We're grateful for every bit of support. Now, back to the hardware. Where do you see this technology going in the next five years?
Lucas: I think we're going to see MEMS mirrors integrated into more than just lidar and AR. Imagine compact projectors in your phone, or tiny spectrometers that scan a laser across a sample. The convergence of MEMS with photonics is going to unlock a lot of new applications.
Lucas: There's also work on using multiple mirrors in an array to create solid-state beam steering without any moving parts, but that's still early. For now, the single mirror approach is the workhorse.
Luna: And what about the integration with AI? I imagine the control loop could use machine learning to compensate for nonlinearities or drift.
Lucas: That's actually happening. Some companies are using neural networks to predict the mirror's response and pre-compensate for errors, which allows them to push the mirror faster and with more accuracy than a traditional PID controller. It's a nice blend of MEMS and software.
Lucas: But there's a cautionary note: the more you rely on software, the more you have to worry about calibration drifting over time. A MEMS mirror is a mechanical device, and its resonance frequency can shift with temperature and age. So the software has to continually adapt.
Luna: So it's a learning control problem, not just a fixed look-up table.
Lucas: Exactly. And that's where I think we'll see a lot of innovation in the next few years. The hardware is getting mature, but the intelligence around it is still being developed.
Luna: It's exciting to watch. I'm curious, for a listener who's an engineer, what's the one thing they should take away from today's episode?
Lucas: I'd say this: MEMS galvos are not just smaller versions of the old galvos. They're a different beast with different trade-offs. If you're designing a laser scanning system, you have to rethink your assumptions about speed, angle, and power. And the good news is, the tools and components are now available to experiment with, even at the hobbyist level.
Luna: That's a good note to end on. I'm looking forward to seeing these in the wild more and more.
Lucas: Same here. And if you want to see one up close, they're often in surprisingly cheap lidar modules these days. Take one apart and you'll find a tiny mirror doing all the work.

Shows are listed from Apple Podcasts' charts and directory, and Podcast Index where it is used. Every show and episode belongs to its publisher and plays from the publisher's own feed; this site hosts none of them.