 I said I make Tesla coils and did this as a hobby purely because all of my other hobbies have generally become my profession and I wanted to keep something that is utterly useless and cannot possibly be Made into any form of business Because otherwise it isn't a hobby The ubiquitous safety talk There are a number of things you've got to be careful of with Tesla coils electrocution is an obvious one certainly the Spark gap one start with 10 kV You then put those up to 100 200,000 volts believe me It's the 10 kV that's more dangerous because that's got current in it. I've got explosion there. You may think that's amusing I have seen transistors explode and they come at quite some force because they are a small container with a lot of power potentially driven by capacitors That also explode when over driven get hot Although to be fair most of them just spit out horrible electrolyte, which is poisonous Burns obvious one electrical burns, but of course the arcs of these things produce are also very very hot Yeah, I've had RF burns hands up anybody who's had an RF burn It is one of the most unpleasant sorts of burns you can get it goes right the way down to the blood And it quarterizes its way all the way through and takes an age to heal up. You don't want them It smells like roast beef. Yeah Pacemakers and cochlear implants if you're doing any form of Tesla coils Don't do it if you have either pacemakers or cochlear implants. I believe the people in this room can work out why So we'll start with the history Classic spark up Tesla coil you start with 240 volts you chuck it into a 10 to 20 kilovolt transformer It charges up a capacitor When it gets to a certain voltage your spark up breaks down Pushing that Charge in the capacitor into a small primary primers on Tesla coils are usually between two and I think the most I've ever seen is 15 turns That goes into a secondary that is a transformer. The secondary has between 200 and 2000 turns By purely a transformer theory That's very small amount of turns. That's a load of turns you get loads of voltage in the top It has nothing to do with the ratio of the turns. It is the ratio of the impedance That becomes important later on when you're driving them especially with transformers with a spark gap Who cares it goes bang you get a big lump of RF wobbling around between the capacitor and the primary That is a tuned circuit. It oscillates most Tesla coils I've seen are between 50 kilohertz and some of the really really small ones about 500 kilohertz So you get a nice RF oscillation that carries on until the Capacitor goes flat. Well the capacitor into the primary back into the capacitor back into there that keeps going around eventually that runs out of power Because the transformer is generally quite high impedance So we get big RF pulse RF pulse as I said is multiplied up and then you have a torus at the top commonly referred to as a top load It's effectively one plate of a capacitor All capacitors have two plates. So one K plate of a capacitor makes no sense whatsoever Everything in the room is the other plate of the capacitor They usually work out between about four puff and some of the really really big ones are hundred to 200 peak of arid Yes, just on a large lump of aluminium. I'm talking that kind of size Dangers as I say high voltage that in your primary circuit is where the high current is Five milliamps across the heart is widely regarded as not being very good for you These transformers I've seen up to hundreds of milliamps Microwave oven transformers are regularly used. They are lethal. I hate them 2.4 kilovolts at about a hundred milliamps Absolutely silent. You can't hear them hiss. You can't hear any corona So they're on the live the dangerous you touch him. You're dead. Not a good combination So we'll start with the first really simple Tesla coil You will have noticed that the spark gap has been replaced by a silicon spark. I'm sorry a transistor It has exactly the same purpose It will explode as soon as you put power. No, hang on It will try and explode as soon as you put power to it because it has exactly three turns on the secondary of Usually quite fat wire It is prevented from exploding by the fact that you get a little bit of inductance into there into that secondary I'll come back to the fact there isn't a top load in a minute Which is driven and there is an LED This isn't my circuit by the way This LED does two things one it glows to show that it is on cool Number two is it rectifies that little bit of feedback Applies it to the base of the transistor in a vain attempt to turn the transistor off If it succeeds all fine thing bursts into oscillation You have big multiplication of voltage and you get a little diddy spark at the top. That's quite poor But it's only running off a 9-volt battery. It's great for amusing kids and It's light fluorescent tubes and that sort of thing Of course if that doesn't turn the transistor off the transistor explodes And then you need to feed more silicon into it and work out why These are very very commonly sold on eBay you can pick them up for 510 pounds something like that and as I say they're a brilliant introduction to solid-state Tesla coiling a Number of their problems. They are low power. I Have seen the scale up of that circuit I've seen it with quite a large MOSFET and mains driven and he had one big problem that he admitted on his site Basically he used to feed it with I think what was sort of a bandolier of transistors for when it Didn't decide to start up because when he plugged it on plugged it in it wasn't a to zero or it was a to zero And it didn't get the feedback you can make up all the stories. Yeah, there's also no smoothing on there Which doesn't help for the initial current rush either So essentially they're only for low power The other problem is their CW by CW means carry a wave. They are on all of the time Now that sounds a little strange until you look at the actual magnetics here because If that is on all of the time, then you've got all of the power all of the time and you're creating a discharge all of the time Which is fine. It gives you a hot discharge What he doesn't do is give you very long sparks because for long sparks what you actually need is a lot of instantaneous change Because it's CW don't get that instantaneous change in fact one of the ways of enhancing this circuit is If you put another transistor across the LED so that you can Switch it on and off. It's like a 555 something like that You can then actually get bigger sparks out of it because you're actually getting a big instantaneous build-up and then nothing and The collapse of that magnetic field in there is actually what you want There's also no storage. I said there was no top load There's no top load because this thing works at an incredibly high frequency It has to to stop the transistor from exploding If that were to work at 70 kilohertz that would take a hell of a current The more top load you put in the lower the frequency constant is of your secondary to everything else So essentially it's not scalable. You cannot get big sparks out of it, which let's face is what we're a laughter So we then go to the next version up, which is a solid-state Tesla coil Yes, SSTC inventive name What we've done we've taken the same circuit we had before but we got it interrupted. We're now using 240 volts AC rectified That's particular on doing that was taking a couple of amps off the mains We've gone to a full bridge, which makes it a lot lot more efficient So we're taking rectified 240 volts 320-odd volts bridge rectifying it so we're getting effectively 640 volts into the primary We're using power MOSFETs The downside of this one is we're actually limited by the resistance of the primary circuit now come back to that in a minute I Don't know how well you can see that I was hoping it would go widescreen. It hasn't This is the circuit. It really is quite simple If you're looking at the other one, it isn't actually a lot different So the single transistor has now been replaced by in this particular version I said it was a bridge and then gave you a circuit with half a bridge You can put the extra two transistors in So you've got a pair of something like 4 pf Sorry. Yeah, 4 pf 50s or something like that 20 amp MOSFETs Found in most things including laser switching supplies Okay That powers the primary the whole lot is Powered by a Gate drive transformer, which is by far the easiest way of driving MOSFETs in the Tesla coil Mostly because it gives you isolation and two is because it gives you all four signals in one assuming using a full bridge Paralleled up drive parallel to up drivers to give you a drive into the gate drive transformer Note any 5 5 5 to give you the interruption. So now we're getting that on on on Which also means we can abuse these a little and possibly take them a little beyond their spec The one that I showed you in that diet that photo. I think was doing about 25 30 amp pulses with 20 amp devices If you can keep it below their current dissipation that works quite well I will come on to later how we can abuse that even more And then we've got a basic amplifier circuit which And this one just says antenna And that quite literally is just a piece of wire stuck up in the middle of nowhere Tapping a little of the energy coming off the top load Capacitively, it's one plate of a capacitor So you get a little spark come out of there. It's picked up by the antennae. It's amplified It drives it into here those two transistors wobble the transformer. You get a bigger spark. It goes round and round around It is thoroughly controlled by the 555 its mark space ratio to give you a decent amount of power without taking a lot as Is most Tesla coils if you don't have that control They are usually to get the biggest sparks out Not ideal for running Constantly by not ideal. I mean usually you expect all the components to be running pulsed So if the pulse has stopped your transistors explode, you'll hear this phrase your transistors explode quite a lot It's kind of the way that the electronic Tesla coils go It's one good thing about a spark gap. You can blow up a spark at multiple times transistors generally give you one chance so Pushing this a little further. This is a colleague of mine's coil in the in in Scotland This is his dual resonant solid state Tesla coil now by dual resonance is what we've actually done is we've added a capacitor in the primary Yes, just that not quite but pretty much just that So we now got a tuned primary Because we have now got a Resonant primary what we can actually do is we can actually feed this with multiple pulses So rather than giving it one. We're now giving it bursts of RF And because that's a resonant circuit and because we're giving it the right kicks We can actually build that like a swing And we're storing some energy now in the primary as well as just in the secondary the magnetic and the top load So we've got a lot more Space for bang coming out of there So I'll go through the things a little more. We've also got a different feedback system We've added in this wonderful thing of current control Now much as I like taking transistors beyond what they're designed to do You can also guess at it till the cows come home We're running them very close now So what we actually do is we're taking some control of it and putting the current in a feedback loop So that we at least know roughly how badly we're abusing the transistors We've gone from MOSFETs to iGBTs purely because they are they're on Resistance is so so much lower and they also have a very fast tail off and d IDT They're basically like an SCR that you can turn off very quickly Ideals for this sort of thing because that mimics that spark gap that we're looking at in the Previous diagram and we've got spark progression now spark progress is a wonderful thing whereby if you make a spark if you then make another spark immediately afterwards The ionized channel is still there So you don't get the spark of the same size you actually get that spark plus that spark and then if you do it again You get another spark on the top So hundred thousand volts will give you about that as a spark actually probably less. Yeah about that So you do that twice you can actually get that and three times you get that four times But you're still only pumping hundred thousand volts in the top So what we've got is a multiplication of what we're doing with no extra electronics No added voltage, you know nothing else and that is a big thing in DRS STC's because that's what gives us the the multiplication from the first coil that was sort of a spark like that To that sort of capability so It's operation We'll start at the the interrupter now you'll notice the interrupter here. I have put On a fiber there's a very good reason for that is generally you are holding that interrupter Being connected to something that's likely to do two or three foot sparks is generally regarded as not a good thing so we separate the user from the equipment by The biggest five and biggest opto isolator we can find which is usually about three or four meters of fiber optic But all that does is produce these sort of pulses Those pulses go into the control And this is an interesting sort of anything in blue is kind of control anything in black is is the actual signal path So the control just literally opens that gate So what that gate does is absolutely nothing because it hasn't got anything yet because nothing's come out of the top So hang on we're starting the pulse it gets a rising edge it creates a pulse a really short start pulse That goes right the way through the phase detect and into the driver as just a pulse pulse goes through gate drive Transformer pulses the bridge That produces a very small spark, but that very small spark Is now detected by this current transformer And all that current transformer is there to do is to feed a phase detect circuit so that we can put the right signals into this chain To drive your bridge at the right frequency and phase So he goes all the way through and then this time we get a much bigger spark because we've got a whole half cycle now And this repeats it does the other half cycle and it will carry on until that control says nope. Sorry. You've had enough They also notice we've now got another current transformer This one's separate to the first because it's being used for a different purpose Which is it is looking for actual current in The bridge in the primary circuit When that is above a pre-detected level it actually tells the control to switch everything off and You also notice very important. I got this capacitor in the drive The problem is you can't just switch transistors off when they're driving a large inductive load and especially a resonant inductive load Because when you switch the bridge off two things happen One is that goes very high impedance and two is the voltage across their Skyrockets now the theory is in the voltage skyrockets that gets multiplied by the transformer and that's actually what produces the sparks That's fine But in overcurrent situation what actually happens is the transistor is at full power you switch it off mid mid Transition and because we're pushing the transistors right to their absolute limit the transistors go well I don't know what to happen this you then get ground bounce you get all sorts wobble on it and usually you lose a transistor So what this control circuit also does is to make sure that it'll only turn off at a zero cross Regardless of whether you're going overcurrent in that cycle or not because a hard transition will blow that transistor This bridge circuit here and in the let me just flip back to that's now going to go more than one. Yeah, that's what it would In that what we're getting in there that particular coil had a thousand amps roughly speaking on its primary coil at 620 volts rectified mains You'll be glad to hear that the transistors here we're using were rated at 200 amps And we'll come back to how we can get away with that in a minute This is the expanded circuit Steve Ward if you ever want to do anything electronic coils it is the go-to reference He was one of the two developers of the DRS STC circuit And this is about the simplest one it can get I've missed the the bridge at that side and the Transistors and they test the coil But essentially we've got exactly the same we've got a drive circuit Actually, we've got two drive circuits a drive circuit driving a more powerful drive circuit, but he's actually now Driving what we call bricks which are Big IGBT's using the rail and car industries that they are physically the size of a brick the big big transistors And then Where are we power supply at the top? You've got interrupter Which I think he's No, there is the optics there So that goes in there's a set of flip-flops that determine whether you are in or out of phase and then there is a signal path from the Where is it feedback? Current transformer that if you see just literally are clamped Amplified and then go straight into a signal path going straight out to a pair of flip-flops and the The drive itself to the transistors so essentially we're back to that very very first circuit of We've got signal coming in being amplified straight out to the coil and straight Straight out to the transistors and straight to the coil He's added in an interrupter the interrupter also determines what phase we're at because again from a manual Control you don't want to switch it off in the middle of a cycle So that's also there and then we've now got the addition of this over current detect, which is a bridge rectifier against the CT transformer with a burden and then is amplified and the set of a Range there to say anything over this. I'm going to flip a signal into this to cut it all off So it's really quite simple Apart from the fact that is quite the simple circuit these days that there's a lot more in there now So I've mentioned about abusing transistors and This is a general sort of transmit Transmission curve from gate to emitter voltage to collector emitter current Now then you'll notice here was there's a five microsecond pulse width The coils we're using five microseconds would be about flat out So actually what we're doing with a running nose at one microsecond sometimes 500 nanoseconds sort of pulses You'll also notice that the gate to emitter voltage only goes up to 10 But the device itself says a gate emitter voltage up to 20 is perfectly acceptable There is also the problem like any transistor. It has gate capacitance So you've got a real bad storm here of large gate capacitance with a very small pulse So we have to drive that gate really really hard and we do this two ways one we put big honking drivers in I've got a little battery powered Tesla coil and that has 14 amp drivers to the MOSFETs and Then the MOSFETs are running at about 300 amps rated at 50 So what we actually do is you go, okay? 10 but we're getting up to 20 and we've got a less of a pulse width I'll get my ruler and pens out And we take that curve and we're going to extend it off this way somewhere and that's roughly where we run it Way off where it says on the data sheet The problem with that is It's absolutely fine As long as you can get the heat out of the device The devices aren't designed to do this So you have to be very very careful when you pick your die and bonding to make sure that you can get the heat out Because we're probably only running at a couple of hundred watts, but with megawatt pulses It's quite normal 300 volts 300 amps and megawatt pulse might only be for one microsecond So That's the way we abuse that transistor that well any of the transistors we do and we do it purely because if we Wanted a thousand amp transistor They are Around four or five hundred pounds each and we want four of them Whereas a 300 amp transistor that we can abuse to a thousand amps is in the region of 20 20 Yeah, 20 30 pounds now something like that And of course we're not running them solidly on Quicker thing about gate drive transformers. That is a gate drive transformer hand wound all of the cores together It literally gives you five windings one in one for every transistor And then you just flip them over to get the right polarities on all of the Bases and then you just literally clamp that just make sure there's no sparks from coming or the way you're gonna get interference from of no idea And talked about the control that the circuit I showed you used General logic So just basically time as an analog. I've now got one that uses microprocessors for most of the control FPGAs are coming in bid style because of their speed and the flexibility of them and because of that you can now do envelope control of the Voltage going into the bridge in the first place and there are ways you can persuade sparks to do some very incredible things That is an entirely different lecture But some really interesting stuff coming out on that The biggest thing with all of this is the overcurrent cutoff timing as I've said it's got to be fast But it's got to be ideally at a zero cross I said in my introduction testicles generally suicidal other than the transistors trying to destroy themselves Primary strikes, which is a spark from there down to there Happen quite regularly the shape of the top load stops up from happening because it actually produces like a magnetic field kind of Static field round the top Tuning is a big one. This coil actually failed. I don't even see the sparks running up and down there. That is not normal That failed because he got a pri a strike to the secondary It disconnected one of his turns on this on the secondary coil and put it out of tune And that's what happened when your coils out of tune is that every every coil Sparks to the next one that's in phase all the way up and down the coil. It looks really pretty makes a superb photograph He was so hacked off Bad feedback is probably the other major one If you don't get a good signal coming back from either your current transformer or from your aerial Then you're gonna fire it at the wrong phase. It will start to oscillate at what that Loop timing is and not at its resonant frequency and then you'll get that sort of stuff as well I run to testicle meet-ups if you're interested one is nottingham gas fest is the one's Cambridge, Tesla Thon I'm gonna be around for a while. Please. Please ask And I want to thank everybody that's ever shared anything on Tesla coils because it's an Arcane art to the internet is made very very accessible and I'm massively thankful for that So brief questions