Transcription
I want to show you this type of levitation that I bet you haven't seen before. Like maybe you've already heard of ultrasonic standing waves, superconductor flux pinning, spin stabilized magnetic suspension, hovercraft, electromagnetic suspension, telekinesis, or simply believing in yourself.
But this type of levitation is different to all of those. And I really like this one because straight away it feels like it just shouldn't work. And even better, it was discovered by accident by Bob Collins when he was asked to investigate why British torpedoes kept missing their target. I found out about this because Bob's son emailed me and that's how we ended up in Bob's front room in Bournemouth and how I, Steve Mold, personally invented a new way to play hockey.
Well, well, tell me the story. I started working with a a young man on a bench assembling parts. This was part of a torpedo project. People had decided that the fault was in the transducers. And I was asked to investigate.
When Bob says transducers, he's talking about ultrasonic transducers. In other words, a piezoelectric device that's a bit like a speaker except it's designed to create ultrasonic sound that can pass through water. These transducers send out pings of ultrasonic sound that reflect off targets. The reflected sound is then picked up by the same transducers. In other words, it's part of the guidance system.
One of the first things I did, I held the transducer in my hand and I happened to have a glass lens on the desk which I placed on the transducer and immediately slid off. Quite mystified by that. I picked that up and tried to keep it on the transducer. I couldn't. And then I placed it down onto a glass paperweight and it started skidding around and that created quite a bit of interest in the lab. A few people gathered around. We were trying to explain what it was. We didn't really understand it.
When you discovered it by accident, were you deliberately a resonance or did you just happen to be? Well, I had that exact signal generator. That is that is the one.
So, this is a signal generator that generates about 30,000 hertz. But these ultrasonic transducers only really get going when you hit their resonating frequency. So when we turn this fine-tuning dial, you can see the moment it hits resonance because the thing starts to move and slide around. Oh. Oh, there you go. Look at that. Oh, that's crazy.
Actually, when I first saw this, I did think about the bristlebot, which isn't really levitation. and it's just repeatedly hitting a surface, which it does reduce friction, but it's not levitation. Bristle bots are really fun, by the way. Like, if you angle the bristles, it actually generates propulsion. But anyway, how do we know that this thing isn't just like a bristlebot? How do we know that it's genuinely levitation? Well, Bob conducted an experiment where the transducer was resting on a metal surface, and that completed a circuit. But when he turned on the ultrasonic generator, the circuit broke. So that proved there was a gap between the transducer and the metal surface. I didn't try and measure the gap, but looking at other people's research, it seems to be about 100 microns. So although it doesn't look as impressive as hoverboarding in terms of sheer numbers, it is technically levitation. And they actually use it to move silicon wafers around in chip factories.
While Bob was showing me his setup, I had an idea. Could I use this to make an airless air hockey table? Though I suppose you wouldn't call it air hockey cuz there isn't an air pump. Ultrasonic hockey. Well, I had a go and it turned out to be a lot harder than I thought it would be.
But first, how does this thing work? Well, you might have heard of ultrasonic levitation before, but there are actually two types. And the type that you'll have seen before is this one. There are ultrasonic sound waves bouncing up and down in this column creating a standing wave. You know standing waves, right? This is my favorite example of a standing wave. It's a bit gratuitous, isn't it? A flame tube. But anyway, you can trap very light objects in that standing wave. But the ultrasonic levitation that Bob just showed me is the other type of ultrasonic levitation. By the way, if you're interested in how any of these other types of levitation work, I'll be touching on a few of them in future videos. So, make sure to subscribe if you're interested in that and maybe click the notification bell.
So, if it's not a bristlebot and it's not being trapped in a standing wave, how the heck is this thing levitating? Okay, so here's what I figured out. When you bring two surfaces together slowly, all the air between those two surfaces gets squeezed out and then they're touching. And that's what happens in our everyday experience. The air just gets out of the way. But what happens when you bring two surfaces together really quickly? Well, if the air doesn't have time to escape, then it gets squished. It gets compressed. Like a tiny bit escapes out of the sides and given enough time, all the air would escape. It might feel a bit counterintuitive that air behaves in this way, but actually you're probably quite familiar with it. For example, if you've ever played with a brand new deck of cards, okay, but what happens if the surface is moving back and forth really, really quickly, like with the ultrasonic transducer, well, then you're bringing the surfaces together really quickly, but then immediately pulling them apart again. So, here's the cycle. The air gets squeezed and a little bit escapes out the sides, and then the air gets stretched and some of it comes back in.
Now, if I was holding the transducer just above the plate, you might expect that cycle to just continue forever. But if I let go, well, all the weight of this thing is now pressing down as well. In other words, with the help of gravity, the downstroke is squeezing the air harder than the upstroke is stretching the air. So the pressure difference on the downstroke is greater than the reverse pressure difference on the upstroke. Actually, there's a second reason why the pressure difference on the downstroke is greater than the pressure difference on the upstroke. And that's related to the fact that the relationship between pressure and volume is nonlinear. I won't get into the details of that, but the point is we should expect that with a larger pressure difference, there should be a greater movement of air. Like when I squeeze air out of this bag, if I squeeze harder, then more air comes out. So if the pressure difference is greater on the downstroke than it is on the upstroke, we should expect more air to be expelled on the downstroke than gets sucked in on the upstroke. So that over time eventually all the air gets pushed out and we're back to bristlebot. But I'm saying that the opposite is true. Less air escapes on the downstroke than is sucked back in on the upstroke. And that's because the size of the opening also matters. So again, when I'm squeezing air out of this bag, less air escapes if I make the hole smaller. And of course, on the downstroke, the gap that the air can escape from gets smaller because the size of the gap is proportional to the distance between the two surfaces. And you might think that these two effects would cancel each other out. You've got increased pressure pushing the air out faster, but then you've got a smaller gap, so not as much air can come out. But what you find is that they don't cancel each other out. And that's because there's another factor that makes it even harder for the air to escape as the gap shrinks. And it's basically drag. Essentially, air doesn't flow smoothly over surfaces. Instead, the surface drags on the air. And so there's always a gradient in speed near a surface like this. That means that when the gap shrinks, you're removing the fastest part of the flow. We don't need to get into the maths of it, but the way it all shakes out is if you half the size of the gap, well, the pressure doubles, but the ability for the air to escape out the gap actually goes down by an eighth. So in total, the flow rate is quartered. So more air gets sucked in on the upstroke than gets pushed out on the downstroke. That means that this ultrasonic transducer is essentially acting like a pump that is pulling air into that gap. And as more and more air gets drawn in, the pressure goes up until that pressure is enough to hold the weight of the transducer and then we're balanced. The amount of air being squeezed out matches the amount of air that's being pulled in. We're in equilibrium and we're levitating.
The animations I've been showing you are actually a 2D gas simulation. I didn't have high hopes, honestly, when I first started coding it up, but all I've got here is a bar-shaped mass that falls under gravity. And amazingly, it seems to work. After the simulation has been running for a little while, the number of particles of gas leaving the gap seems to be the same as the number of particles that come back in. And so the bar levitates.
But what about my idea for a variation on air hockey? Well, here's my thinking. Turn that ultrasonic transducer upside down, stick a metal plate to it, laser cut some little acrylic discs, and you've got yourself a miniature game of hockey. It's a bit like air hockey except you don't need a pump to be pumping air. Everything just levitates because of the ultrasound. But it turns out it's not that simple. For a start, how do you go about buying torpedo parts without ending up on a list somewhere and actually I couldn't find any transducers like that anywhere. The closest thing I could find is the type of transducer that's used in cleaning baths. But like why can't I just use a normal speaker? Why do I have to use a piezoelectric speaker? In fact, speakers are so ubiquitous. How come we haven't seen this type of levitation loads of times before? Well, it turns out it's to do with something called impedance matching, which I made a whole video about. Link in the description if you want to find out more. But basically, the issue is this. On the downstroke, the pressure really spikes because it's inversely proportional to the volume. And a normal speaker only has enough muscle to push air back and forth at standard atmospheric pressure. So it would fail on the downstroke. Whereas a piezoelectric transducer is strong enough to push water back and forth. And water is 800 times more dense than air. In other words, the difference between a normal speaker and a piezoelectric speaker is like the difference between a human voice box and a dolphin voice box. That's why humans can't be heard underwater and presumably why dolphins can't be heard on land and ultimately that's why we can't talk to dolphins.
Actually, dolphin anatomy is really interesting, but it's probably a tangent too far even for me. But honestly, I could talk about dolphins all day because I just love animals. Actually, if you're anything like me, it'll probably make you quite sad to know that a lot of dolphins have to shout to be heard over the noise of the engines of tourist boats. I mean, look, there's a lot to be sad about when it comes to the impact that humans are having on the natural world, which is why I joined Planet Wild, and it's why I'm partnering with them for this video. Because although the challenge is great, we don't have to feel hopeless. Every month, Planet Wild's community of over 20,000 members, including me, funds a mission to bring back endangered species, protect our oceans, or restore forests. By the way, if you're interested in biology videos, Planet Wild has a weirdly good YouTube channel. But actually, it's an important part of what they do because every month we get to see our money at work because they make a documentary about it and post it on their channel. For example, a couple of months ago, they funded special dolphin microphones to figure out how to help them communicate without having to shout. Piezoelectric microphones, of course, because of, you know, the the thing we were talking about. Planet Wild doesn't just highlight the problems, it funds the solutions, which is why I became a member. If you want to become part of a growing community that makes a difference, consider joining Planet Wild. You can give any amount, big or small, whatever feels right to you. And the first 150 people to sign up using code mold 4 will get their first month paid for by me. Just scan this QR code or click the link in the description. And after that, you can cancel anytime. And if you want to see them in action, check out their Dolphin Mission video. Link to that in the description.
Right. So, you can't buy torpedo transducers, but you can buy transducers for ultrasonic cleaning baths. The type that vibrates a tub of water for cleaning jewelry and stuff. They look like this. They're a bit heavier than the torpedo transducer that Bob showed me and it has a slightly different resonating frequency, but maybe we can get something working with this. Figuring out how to drive one of these things was also a bit of a nightmare. Like I considered using the circuitry from inside one of those ultrasonic bath cleaners, but then you can't specify the exact frequency with those and I needed to be able to do that because of the resonance of the hockey table. So I looked at more custom circuitry, but then the transducer, you know, it's actually a capacitor as well, and when it hits resonance, all its characteristics change. So you can't just use a standard amp that you might use for like audio. You can end up with voltage ringing and current spikes and your circuit and you destroy your transducer. Anyway, I eventually figured out that I could use a motor driver. And I hate to admit it, but I did get some help from that scoundrel electroboom. He's probably going to make a video about transducer circuits soon, so be sure to subscribe if you want to catch that.
The other big problem is you can't just vibrate the plate up and down because it doesn't just move up and down. It wobbles instead. In other words, you end up with Chladni figures. Link in the description for two of my previous videos on Chladni figures if you want to find out more about those. But the point is the salt moves to the parts of the plate that aren't moving. What this tells us is that if you drive a plate like this at 40 kHz, you end up with quite a short wavelength. And if the wavelength is shorter than the width of the plate, then you end up with standing waves. There are points on the plate that aren't moving. These node lines shown with the sand. And of course, if one of our acrylic pucks ended up on one of these node lines, well, we wouldn't get any levitation.
So, what's the solution? Well, instead of standing waves, we want to try and get traveling waves. And to get traveling waves, we need to drive both ends of this plate with ultrasonic transducers. We want to drive both ends at the same frequency, but we don't want them to be in phase with each other like this. Instead, we want one of them to lag behind the other by a quarter of a wave, which looks like this. Not this. This like I think you still get standing waves and node lines and Chladni figures, but layered on top of that is this much larger traveling wave. So, you end up with no dead spots.
So, here's the final setup. One transducer at each end and my signal generator can output two separate signals linked to each other. So they have exactly the same frequency, but one lags behind the other by a quarter of a wave. I've added walls using acrylic, which is very light, so it shouldn't change the resonant characteristics of the plate. And look, there's holes here and here for the goals. How cool is that, though? A miniature game of hockey using ultrasonic levitation. That's crazy, isn't it? We've gone from broken torpedoes to frankly a sporting revolution.
But hold on. What was wrong with those torpedoes in the end? Was it the transducers? Well, I never found out. I think it might be classified. But I did learn something interesting about dolphins that might be relevant. They have a special organ in their heads called a melon. And the melon's made of a special kind of fat that matches the acoustic impedance of seawater to the acoustic impedance of their vocal anatomy. So maybe these broken torpedoes just needed a melon.
Actually, speaking of which, the first 150 people to sign up to Planet Wild using code mold 4 will get their first month paid for by me. Just scan this QR code or click the link in the description. And if you want to see Planet Wild in action, check out their dolphin microphone video here.