Video Transcript
A he new helion fusion video. We're going to watch it. I haven't watched it yet, so I hope it's good. Watch party. Everyone get your popcorn. Get your candy. Here we go. >> So, Helon operates a pulse fusion machine. We have a set of electromagnets. Sometimes we call them the coils that are plugged into a pulse power capacitor bank. This is capacitors charged somewhere in the 40 30 kilovolts range discharging in the tens of kilas each. Uh the coils will wrap a vacuum vessel. We inject a neutral gas into this vacuum vessel and then discharge the pulse power bank into the coils. This will ionize the gas and shape the resultant plasma into an F FRC, a field reverse configuration, which I think of as like a semi-stable plasma arrangement with closed magnetic field lines. And you can move it around in space by applying magnetic field externally. So we make two of these uh one on each side of the machine. We move them towards each other into the middle of the machine where we compress them. As they compress, they get hotter and hotter. Fusion starts to occur. Their internal energy increases and they try to push back on the magnetic field. And that push back creates energy that we can couple back into the electromagnetic system. And we see it as voltage on the capacitors. My team, uh, high voltage and low voltage team, they designed electrical circuits. >> So, this is really cool, right? I mean, we're only one minute in, and he just explained it. They're taking two plasma rings, having them collide, and then we're having them spin around, and then the magnetic field, we we have this p this like push out effect, and the magnetic field compresses them, and we use that to produce our energy. So now we've created like a I don't know how to explain it like a plasma magnetic plasma turbine. It's just going to start spinning and we can just use that to produce energy. That's just how any all forms of energy work. We're just spinning uh a copper coil through a magnetic field. This is the more advanced version, right? Um so here we go. Let's >> the machine. The high voltage team focuses a lot on the pulse power system. That's the system that delivers the energy and capacitor bank to the electromagnets. The low voltage team works on things like control systems, health monitoring of the pulse power bank, plasma performance monitoring, measurement of energy, measurement of energy, you know, where it's all moving around. They also follow the life cycle of an engineer. They design things. They simulate their hardware. They make prototypes. >> The basics of the pulse power circuit is really something from E101. You have an RLC circuit. You have this capacitor, switch, an inductor, and a resistor. So, you can see those elements in Ursa. You walk into the building, you'll see the capacitor bank, which is kind of the capacitor and the switch. We have the capacitor bank divided up into what's called pallets. These pallets contain a few capacitors, a few switches. It's way safer, way more manufacturable than making one gigantic capacitor. You take these capacitors, you put them in parallel with each other, so they can all flow current together, and they will discharge into cables that are connected to the coils. Cables I think of as our resistive elements, the R of the circuit. There are other elements with losses, but cables are really the only one that's mostly a loss. Uh, and then you have the coils, which are the inductor. Of course, bills have some losses as well. Capacitor bank itself has some. And so, you start to see this in the building as you walk the pulse power bank racking. It's very geometric. You're very much trying to fit components together without creating new arcing uh potentials. >> Just like with magnetic motors, the battle isn't the front end. The battle is the back end. The battle is how do you set up the capacitors in the back end and how do you get the energy to flow to them. And what he's saying here is that we can basically get the energy to flow to our capacitors at near perfect efficiency. And so we can get this super high efficiency where we're having our plasma charge all these capacitors and then we're going to selectively discharge them. Discharge them to use that energy to power the lights or whatever to power the power grid. And then we're going to have them keep recharging. Right? So, not only do we have to have our fusion reactor running, but we have to move that energy from our fusion reactor, our electric fusion reactor to our batteries and then remember like if I just have the system open, then it's only going to recharge to a certain level, right? So, we need this battery over here to start charging up and then we need to switch over to a new battery, have this one start charging up, and we have to have this one start draining without impacting this one. So there has to be a very complex system of the batteries charging up and then draining in order to use that power actively and efficiently. >> Way I think about it is we are taking a set of simple components and turning up the operating voltages turning up the operating currents. So everyone's familiar with u LC circuits sometimes people call them resonators. Uh the idea for polaris machines is we will take some simplified version of that and dramatically increase the operating voltage for the target of creating more magnetic field which is a function of how much current is flowing. The the basic scale exactly the way you'd expect but now you're forced to think a little bit more about what exactly happens when I when I tell my switch to go into conduction. What exactly happens when the switch comes out of conduction? How are things coupling to each other? But at the end of the day it's really a scale up of an otherwise relatively simple power electronic circuit. Specifically at Helon, we're driving gigantic electromagnets to try and generate enough magnetic force to form and translate plasmas to then do fusion. And that requires uh some really high current pulses, different pulse widths, um lots of cool circuits that have to do that and drive circuitry. And all of that falls under the umbrella of pulse power. We have a huge capacitor bank that has to drive high voltages through our semiconductors. And um and then they actually will see reversal of voltage because of the way that an inductor works. And so it is purely an electric to magnetic conversion. They must then receive that negative voltage in a way that we can then either capture it or hopefully one day put back on the grid. We design our units to be manufacturable in what we >> negative voltage. This reminds me of Deborah Chung. Why does this remind me of Deborah Chung? So they're saying, "Okay, we put our we put our voltage to charge up our plasma generator and then the energy comes back to us. Now we need to take that energy and we need to go use that excess energy that we gained, that overunity energy that we gained. We got to go use it to charge these batteries and then we got to start the process over again." And as long as you have less energy coming into this system than you're getting out of this system, you've got unlimited green energy. Then it's just a matter of scaling it up. So they've already proven conceptually that it's possible. They've already proven conceptual level obviously even to get to this point where they're already building these giant things. They've already proved it conceptually. Now they're actually just building plants. They're actually building and they're saying, "Okay, we're going to make one of this size and if it's this size, it can produce this much energy and then that energy will be added to the grid." Or what's really happening, they're going to power Microsoft's AI data centers. Right? That's what they're going to do. They're gonna power Microsoft's AI data centers. they already have a contract to do it >> called pallets. Uh and that means they're just subunits of the same circuit that are then run um in series or parallel and there's more series and parallel in the pallets themselves. Uh so that means uh different methods to make sure that everything is balanced in both their turn on and their turn off so that uh you get a uniform energy and field delivered to the coils. So I built in this bay all the hardware necessary to recreate the fields that a formation coil would see. uh it carries less than 200 kg which is a lot uh and we're trying to understand more deeply how these circuits interface directly with the coil. We believe that uh to make the best plasmas we need to have as high flux as possible in the formation section which means you have to have incredibly fast and high formation loop voltage. And so that for us that means the uh the semiconductors the circuit needs to be more balanced and uh steal less voltage out of the capacitors. We were connecting these all to a formation coil. This does mean potentially uh more robust systems and learning more about the geometry potentially of how we connect to our coils. So, as you can see, we've got quite a few pallets all hooked up running their little racetrack all the way through and then they come and connect to our coils. This setup is using the exact coils that Polaris uses in formation section. Same hardware, same size and so that we can hopefully create the same field and we can learn from that. Uh so, we then also had to replicate how they're connected. This setup will vet that that is the right connection method to do as well as help us understand if the way that we've connected them has created another loop. The real question is is that loop then also >> you guys see those electromagnets? Those are electromagnets, right? Look at those electromagnets. That looks just like many free energy magnetic motors. Look at that. Look at how they're configured geometrically around their uh circular uh stuff there. This looks just like the free energy magnetic motors using electromagnets. Very similar, right? And you're seeing the same concept because we've basically taken the free energy magnetic motors from the 70s and they found a better way to do it. They said, "Why bother with these magnetic motors that are big, they have moving parts. Instead, let's just use plasma. Let's go straight to the source. Plasma is the source, right? This is why, you know, free energy magnetic motors are cool. They work. They're obviously real because 0 point energy is real. But this is taking it to the next level. This is why I say that fusion fusion is basically just um a tap. It's a it's a zero point energy tap. So a magnetic motor is basically the analogy or the equivalent of like a watermill which is like yeah watermill is cool. It works right. It it captures the stream of water and uses that to produce energy but it's not really efficient. It's not really efficient. If we go right to the source, then we're taking plasma and we're using plasma itself to do it. Now we're getting way more efficient. >> And if the way that we've connected them has created another loop, the real question is is that loop then also reducing the performance of the machine. There are voltage probes in different areas. There are current probes in different areas and we are now using this to vet some new plasma diagnostics. >> The post power bank is made up of tens of thousands of semiconductors and somewhere in the thousands of capacitors. Is it performing the way you expect? Is it turning on when you want it to? Is it stopping when you want it to? That's information that in a pulse power system of this scale, you typically would not have had access to until something's broken. You have to go look at it out there. >> Hello. >> We had a large problem, >> a gap in how do we go collect a bunch of data off of a highly distributed machine, a complex 3D array of of different points where we wanted to take data >> and then that same complex 3D array go and trigger all of these in a specific sequence. The two choices that we had to wrestle with was the choice between a centralized system and a distributed system. >> Okay, now with >> I actually understand what he's talking about here, chat. Now I understand what he's talking about here. So now he's basically explaining how they're going to make the energy distribute. So they're building this big system of batteries. This now now you're in my wheelhouse databases. Now you've joined Ashton's wheelhouse. So now they're saying, "Okay, how are we going to get these batteries to turn? How are we going to get the power to flow through this system? How are we going to get to pull from our reactor and charge up and then distribute to the next one next to it and go down the line until it gets flowed out to the public to my house so they can charge my sign so they can power our our, you know, internet and what have you while it's running. We need this thing to be running all the time. So, we have to have a complex way for when we're going to turn on or turn off these batteries that we're making. >> And you would have all of these tendril harnesses worming its way through this whole complex 3D array. Probably more over here, too. And then, you know, even more in between those. What we didn't like about this one is the uh ability for redundancy either now or in the future. Is there a pathway to make this a redundant thing? >> I love this. to make it redundant. You have now probably very expensive electronics that are here and you have to go do two of them, maybe three of them. Does that mean you also have to double or triple your harnesses? >> Okay, so what he's saying is there's two ways you can do this. One, you've got your reactor over here. You can connect your reactor to every single battery, right? We can either connect it to every single battery. That's decentralized system. It's decentralized, meaning it the reactor is connected to all the different batteries. And then you turn off or on whichever battery, right? You say, "Okay, fill up battery number one. Now turn that off. start filling up battery number two, have battery number one start shooting that energy out. Decentralized centralized system would say, okay, send the energy to one system and then start distributing it. Have the one system start distributing it between the various different batteries. And of course, there's positives and negatives on both different systems. That's what he's explaining. And the third thing that makes it really difficult is the integration and the maintenance of that system. During operations, if you pinch a harness by accident, well, you've broken it, maybe you'll have to go run another one. For a distributed system, what you'll end up having is smaller compute nodes >> all throughout, >> connect them together. >> Okay. >> From a redundancy story, you can close uh uh uh it might be more cost effective to do, you know, redundant smaller controllers and also maybe redundant harnessing this way. And if one breaks, you can go and just put another one in there quickly. And so it's a more dynamic system for uh uh our ability to measure diagnostics off the machine and help trigger the machine. We chose to do a distributed >> distributed distributed 100%. Right? Because distributed is modular. So now you say, "Oh, if one of my tendrils breaks off, okay, that's fine. Just redistribute it to the other ones and then we'll replace the one tendril and then we we're we're back in business. If we need to add more, then we can just add more to one of our distributed networks. We've got six distributed networks. We need to add more. We add more to one of them." Boom. Each of these uh uh independent uh compute nodes does look like the graph >> take local telemetry do local actions flag local flags all to a uh you know a smaller more flexible control system that is cheaper and faster to deploy than something more centralized. That's what we have today. We wanted to measure things in the single volts range or maybe a few kilovolts range but on top of a common mode voltage of tens of kilovolts. How do you do that? There are consumer offtheshelf like probes that allow you to do that but not at the scale that we needed for this. >> Another challenge was the circuitry to go make those measurements requires some small amount of power. It is not super practical to take from the high voltage part of the system and buck down and provide a few watts of power for some 5 volt or 3.3 volt device. It's not a practical thing. It would require maybe a lot of exotic circuitry that we couldn't necessarily scale or just an incredibly large amount of inefficiency and power burn. So we had an idea what if we look at wireless charging and so we thought okay we could take a primary side coil which is ground reference and we could take a secondary side coil that would be you know enabling some piece of you know future electronics to go float and an insulator in between that can withstand common mode voltage that we were expecting tens of kilovolts both from a punch through perspective and a tracking perspective. This early unlock allowed us to scale thousands of measurements and let me show you what it ended up looking like. >> Wait what? These are wireless power transfer you know devices. We have a primary side which is that primary side ground reference coil and we have a secondary side which is that side that will float >> and in the middle we have a insulator material that we designed both from a punch through aspect and a tracking aspect. >> Now we can transfer a few watts of power very efficiently over a galvanic isolation boundary and power whatever we'd like. On this board we have a >> Okay. So what is he saying here for the people? Basically, he's saying, "Okay, we've got our coil on this side. We've got an insulator in the middle, and we've got another coil." So, these two coils, the anode and the cathode, they want to have current flow between them. They want to have current flow between them, but they can't. They can't because there's a there's an insulator between them. But when the current gets high, when the voltage gets high enough, just like with lightning in the sky, why does lightning strike? The voltage builds up. And when the voltage builds up high enough, it can break through the insulator, which is the air. The air is the insulator and therefore the lightning strikes the ground. So they're using this same concept of wireless energy. They're saying, "Okay, when the voltage builds up high enough, it'll break through the insulator." And this is how we can control how much energy we transmit. So only when the voltage gets to a certain level, then we'll have it automatically transmit to this other location that we want it to go to. Oo, spicy EDC in a microcontroller to go digitize some analog measurement that we're curious about and go telemetry it out with fiber optics on this. Just kind of a silly prototype, but we wanted a Raspberry Pi that could float to also tens of kilovolts. Well, we could just go ahead and put that on here. What this allowed us to do is scale thousands of measurements all in real time delivering back to engineers and operators to let us know what the health of the machine is, >> what the performance of the machine is, and allow us to measure energy recovered back on our capacitor banks in a very accurate and precise way from Fusion. >> Great video, man. The Helen Fusion has the best videos, chat. I'm all in on the Helen Fusion videos. Best videos. They explain everything and I feel like they might regret explaining so much earlier on because I don't think they expected someone like us to come along and be like, "You guys have it figured out. We caught you. We It's like we caught them red-handed.