- Blog entry posted in 'Building a Dual-Resonant Solid State Tesla Coil', August 20, 2014.
Hi everyone, welcome back to my blog!
I lied when I said the next entry would be about the build. I left out something very important--the final schematic! I discussed the different parts, but never explained how they go together! This particular schematic is essentially one of Steve Ward's DRSSTC designs with a couple of tweaks for my coil. I cannot take full credit for the design.
Since my last post there have been many changes, the most notable of which being the addition of a flip-flop to allow soft-switching of the IGBTs (only turn them off during the zero-crossing point of the current), and of over-current protection. Let's start by taking a look at the full schematic. Please note that this circuit is subject to change--there are always things that will need to be adjusted throughout the design process.
Pretty confusing, huh? I think so too. So let's break it down into its key pieces.
1) The Interrupter
I didn't actually include this in the schematic because there is any number of different ways to interrupt the coil. Generally, depending on your coil, you don't want the on-time to exceed 100uS, and generally you'll want even less. I believe I attached a calculator program written by Sigurthr (over at 4hv.org) in one of my previous posts. If you enter your coil's information, it will calculate the maximum burst length for you. Possible interrupter ideas range from simple single 555 timer circuits to complicated ones that require 3 555s, or even microcontroller-based ones. Regardless of the interrupter, however, there is one thing you should pay close attention to, and that's the optocoupler shown in the schematic:
Many people replace this with an optic-fiber cable, which is certainly ideal and offers significantly more protection for the operator and potentially sensitive interrupter circuitry.
2) Soft-Switching Circuitry
As mentioned before, "soft-switching" means that we're only going to turn off our IGBTs at the zero-crossing point of the current. Trying to turn off IGBTs when there are hundreds of amps flowing through them puts a lot of stress on them and can cause them to fail rather quickly. Soft-switching only allows them to turn off when there is little or no current flowing through them. One of the most popular ways to do this is using a positive edge-triggered JK flip-flop such as the 74HC109.
This particular circuit also incorporates the interrupter signal, the feedback signal, and the over-current detection signal so that all of them work together to turn off the IGBTs at just the right time.
3) Driver Chips
The driver chips are simply two paralleled UCC27425 MOSFET gate drivers. Each chip has one inverting and one non-inverting driver, which is required in order to generate a bipolar square wave to drive the gate drive transformer. Some designs use two separate chips such as the TC4429 and TC4420, one inverting and one non-inverting respectively. I chose the UCC27425 because it has both in one package, but failed to notice that it is only rated for 4 amps. I decided to use two of them in parallel to help share the load. Generally the higher the current rating, the better.
You'll notice on the output I have a capacitor in parallel with a resistor. The capacitor is used to block DC which could cause the circuit to latch up, which could lead to excessive heating. However, when you have a capacitor in the same loop as an inductor (the primary of the GDT), you will get "ringing", or unwanted oscillations. In order to avoid this, we add the resistor to damp these oscillations which will ensure we only get a clean signal going to the GDT.
4) The GDT
The "Gate Drive Transformer", or "GDT" for short, is used to drive the gates of the IGBTs. It serves as an isolation transformer which prevents nasty kickback into our driver circuitry, and unwanted signals from going into the IGBTs. The GDT serves several purposes, but isolation is probably one of the most important. In this schematic I use two cores, but I will probably consolidate it onto a single core soon. It will depend on how your IGBTs are laid out whether you'll want one GDT or two.
Notice that there are two pairs of secondaries. Each pair contains two secondaries that are in-phase with each other, but that the two pairs are out-of-phase with each other. If you recall the animation for the H-bridge that I made for one of the previous entries, two diagonally-opposite transistors need to be on and the other two should be off. We achieve this by using opposite-phased pairs of secondaries on the GDT. If you use a half-bridge you will only need two secondaries, each out-of-phase with the other.
5) H-Bridge (or "Full Bridge")
The H-bridge is what generates an alternating current from a steady DC bus. This AC signal is critical in order to get the proper oscillation from your primary tank circuit. The 3300uF capacitor is for smoothing, since the input to the bridge will be rectified mains (120VAC in the US). In order to find the DC voltage, simply multiply the RMS AC voltage (120) by the square root of 2 (1.414), and that means my DC bus will be around 170V. However, a large smoothing capacitor will be very important, otherwise we will get unwanted oscillation and poor overall performance from our coil.
The tank cap (MMC) and primary coil will create a second-order LC circuit that will give us a resonant frequency determined by:
where F is the resonant frequency in Hertz, L is the inductance of the primary coil in Henrys, and C is the capacitance of the MMC in Farads. The trick is to match this frequency with the resonant frequency of your secondary coil and topload. Once you do that, you have a "tuned" Tesla coil.
6) The Tesla Coil
The Tesla coil itself consists of the primary coil, the secondary coil, and the topload. This part is fairly straightforward. Something that should be noted, however, is that there are two current transformers (or "CTs") that are placed on the primary wire. In the schematic it shows two transformers labeled 1:1000, but really they're just toroids that the primary wire passes through. One toroid is for the feedback and one is for over-current detection.
7) Feedback Circuitry
The feedback circuitry begins with one of the CTs that the primary wire passes through. This induces a current in the secondaries of the CT which we clean up using some diode clipping and a Schmitt-Trigger inverter to ensure we get a nice square wave. This signal is passed back into the soft-switching circuitry, and then fed back into the IGBT drivers. This creates a feedback loop which ensures a clean oscillation on the primary.
One important thing to note in the feedback circuitry is the variable inductor in series with a 5 ohm resistor across the output of the transformer. This creates what is called "phase lead", which allows the IGBTs to handle the higher frequencies. Recall that we use the flip-flop circuitry to soft-switch the IGBTs. However, large brick-type IGBTs generally take a few hundred nanoseconds to switch. While this may not sound like much, the rise and fall times add up and cause the IGBTs to switch too late, which leads to significant heating especially at higher frequencies. The phase lead allows the IGBTs to switch slightly sooner than they would normally, to ensure that they actually turn off at the zero-crossing point. By using a variable inductor it is possible to adjust the phase lead to the best possible amount of time.
8) Over-Current Detection (or "OCD")
Over-current detection is used to make sure you don't stress your MMC and IGBTs past their operating limits. One of the best ways to do this is to use a current transformer on the primary coil, and a comparator to shut the coil off if the sensed current exceeds the specified level. I based mine on one of Steve Ward's designs which uses an LM311 comparator that feeds a signal into a one-shot 555 timer. The timer disables the driver for a millisecond or two to allow it to cool down before restarting again.
If you're familiar with comparators and 555 timers, it will be easy for you to follow how this works. When the current exceeds the threshold, a falling-edge signal is sent to the 555 timer's trigger pin. This causes the output of the 555 to go high for a certain amount of time based on the values of C6 and R11. In our case, we use a 10k resistor and a 0.1uF capacitor. Based on the formula for a one-shot 555 ( T = 1.1 * R * C), these values give me an on-time of 1.1mS. While the 555 output is high, it pulls the input of the inverter low through a transistor, and the "clr" input of the flip-flop is held high. Since we're tapping off of Q/ ("NOT Q"), then the enable pins of the drivers which are connected to Q/ are held low for that 1.1mS, which means the coil stays off for that amount of time.
That's about it. Sure there are a lot of parts, but if you break it up into the individual sections it's not so bad, is it?
Anyway, now that we've looked at the final schematic, next time we can start following the actual build. Please remember that at this point in time there will likely be long delays between posts as I tweak things and make adjustments to the individual parts. I've already gone through several designs, so be warned that chances are nothing I have done is final!
I hope you've enjoyed this look at the overall schematic for a DRSSTC. Your first look at the schematic is always the most overwhelming, but it really helps to know how to break it down into the individual pieces.
As always, I welcome comments, questions, feedback, and suggestions. Feel free to post as a response to this entry, or to send me a PM on the forums.
fezder, August 20, 2014
Hey! i read little bit, and i have question: at section 2, soft-switching, is it same consept as with ZVS? Or are they completeley different league? (No worries, i'm not going to build teslas brobadly looooong time, too hazardous for me....)
DerStrom8, August 20, 2014
Hi Fezder, Yes, the idea is effectively the same, except we're working with current instead of voltage. If you take a ZVS (Zero-Voltage Switching) driver, it switches when the VOLTAGE crosses zero. ZCS (Zero-Current Switching) does the same thing, but switches when the CURRENT crosses zero. So, whenever you see "soft switching" in this post, you can also think of it as ZCS, because the current transformer senses when the current crosses zero and switches the IGBTs at that time.
fezder, August 21, 2014
AH i see, can you tell when pros/cons of ZCS/ZVS? i'we seen ZVS in like induction heaters, but never ever ZCS anywhere....
DerStrom8, August 21, 2014
It all depends on your application. In the case of this Tesla coil, I want to turn on/off the IGBTs when the current is zero. We don't really care what the voltage is doing. But some circuits look at the zero-crossing point of the voltage instead. If your circuit is purely resistive, then the voltage crosses zero at the same point that the current crosses zero, so ZVS and ZCS have essentially the same effect. Once again, it really depends on what you're doing.
fezder, August 31, 2014
this doesn't really attend to this blog, but i was curious and i measured switching of fet/fets current, and in my test i build, noisy and all, current was around 2 amps when fet switches off. It doesn't get hot, only thing that gets hot is power resistor as load, so no point getting experiencing with ZCS/ZVS. BUT, i was thinking it would be wise to start experiencing early on and get understanging how these work, so i was wondering as i don't have any j-k F/F on hand, only normal logics and positive edge D-F/F's, could any of those be used in testing? I can't find any help on google either...only those inductor heaters which use ZVS Also, voltage accross 0.1ohm resitor at 2 amps is only 0.2v, so perhaps it's too small for any sensing purposes? Sorry, i can also ask around forum for enlightenment so i don't litter your blog with stupid questions further :)
DerStrom8, September 05, 2014
Hi Fezder, sorry it took so long to respond. Generally you only need to worry about ZCS/ZVS when you're switching LOTS of current at a very high frequency. Obviously your transistors' specs ultimately decide this. The transistors I'm using will probably see a peak of around 500A, so trying to switch them off when they have 500A passing through them is a very bad idea, and is a good way to damage them. That's why I will be using ZCS for my DRSSTC.
fezder, September 05, 2014
No worries, of course you have other things to do also than lurk in forum. Thanks for reply nevertheless!
blu.eyes, April 06, 2017
Hi, i have a question, You have used the 74HC109 flipflop, with a logic circuit which doesn't allow clear and preset to be the same logic value, according the datasheet, the output will be always equal to the input, regardless of the clock, so the soft-switching will not work. I'm right? Can you explain me how it work? Sorry for my bad english, i'm italian.
DerStrom8, April 07, 2017
Hi blu.eyes, thank you for the question -- It is a good one. You would be correct about the clock being ignored if the RC delay made up of R9 and C4 wasn't there. When the interrupter switches off a LOW signal is applied to the CLR pin. At the same time a HIGH signal is passed into the RC delay from IC1A, so the input to IC1B remains low for a period of time determined by the RC time constant and the hysteresis of the Schmitt trigger. This means PRE is also held HIGH for the same period of time. When PRE and CLR are both HIGH, and the J and K inputs are both HIGH, then the output (Q/) is switched LOW only on the next rising clock edge, per the truth table in the datasheet. I hope this clears up some of the confusion.
blu.eyes, April 08, 2017
Thanks, i have considered the r9 and c4 as passive high-pass filter. I have other question Near the CT there'is a inductor whit in series a resistor, how it work? Next there are 2 zener diode and 2 schotky diode, are this diode limiting the max voltage of the current transformer? The capacitor next to the diodes is a decoupling capacitor,a capacitor for a delay,or what? Regarding the soft switching, it's necessary only when the interrupter will disable the bridge or it will be usefull also when the bridge is switched on by a new interruper signal? Sorry if i'm annoying you, but i want to know how a circuit work before i build it. Thank in advance for your reply.
DerStrom8, April 08, 2017
Hi blu.eyes, R9 and C4 actually technically form a LOW pass filter, not a HIGH pass, but in this case they are actually used as a delay due to the relatively high component values used. As for the inductor in series with the resistor, this was actually a part that I removed for the final design. The original intent of the inductor and resistor was to act as a burden that provided what's called PHASE LEAD. This allows the circuit to start switching off the transistors slightly sooner to make sure that by the time they fully shut off, the current is at its zero-crossing point (meaning less stress on the transistors). However, with the zener and schottky diodes in the circuit, the output waveform was actually being clipped and the phase lead was rendered useless. In my particular setup I realized that the phase lead really wasn't necessary, so I just removed the inductor and the resistor from the circuit altogether. The back-to-back schottky and zener diodes are indeed used to clip the output waveform from the current transformer (a high amplitude sine wave) to a +/- 5 volt square wave. Then D9 and D10 clamp the signal to 0 volts and 5 volts which the 74-series logic can handle. C2 is an AC-coupling capacitor (DC-blocking capacitor) that prevents any DC from passing through. This prevents the Schmitt trigger inverters from latching up if any DC is present. It makes sure that the only signal that is passed through and cleaned up is the constantly changing signal coming from the feedback transformer. Soft-switching is most important when the interrupter turns OFF, because it prevents them from trying to switch off when current is already passing through them. If they are turned off suddenly when hundreds of amps are passing through them, it can cause massive voltage spikes due to the inductance of the primary coil. These voltage spikes can destroy the gate oxide layer, ruining the IGBT permanently. You won't get voltage spikes nearly this bad during the turn-on of the interrupter. I suggest reading my full PDF on this design and build. It is much of the same information as found here on the blog, but I have made edits that clean it up and make it more understandable. It is a work-in-progress, but it should be an improvement over this blog: [URL]http://teslaunderground.com/Matt/VermonsterMiniBuild.pdf[/URL]
blu.eyes, April 08, 2017
Thank a lot for the information. Thank for you disponibility. I will read the pdf. Have a nice day.