My Tesla Coil Project

[Einstein]

On your PC Board are you using a ground plane? Also remember the Miller effect.

No ground plane at all on my PC board. I'm still checking out the Miller effect. So far from what I've read on it suggests this is not a contributing factor.

The oscillator is feedback controlled. But other than that the main power MosFet is the only part that actually keeps the oscillation going. So basically the Tesla coil in this configuration is just a basic oscillator. Although using a power MosFet does serve a dual purpose in that the oscillator is also an amplifier as well. I am suspecting that this configuration is also similar to a non-inverting op-amp.

I did move the power MosFet off the PC board yesterday to see if there was any improvement in start-up operation. It will work in continuous mode but not in on-off pulsed mode. But I did notice the voltage swing on the field voltage dropped from +/-32 volts down to +/- 23 volts.

There is something I might mention. You'll notice in the schematic I have the secondary coil connected directly to the input pin of the MosFet driver. With two fast diodes clamping the voltage swing to just 12 volts of the power supply. The diodes clip the input voltage into a square wave. And also prevent any over-voltage condition to the MosFet driver. With the addition of two capacitors at the input I was able to get this to work perfectly on the bread board. I think this allows the secondary coil to continue to ring after turning off the power MosFet. On the scope, I could see this ringing take a while before dissipating. I believe it's this ringing that allows the coil to turn back on during on-off-on-off operation.

But this didn't work on a PC board. The way I got it to work in just continuous mode was to add a capacitor and resistor in parallel with the power side diode. And I did get it to intermittently work in on-off-on-off mode by holding my hand close to the wire connection during start-up.

To me this is all trial and error. I'm going to explore the op-amp angle today. And also Ironcladmarmallows suggestion on adding resistance to see if I can get the ringing oscillation to slow down.

 

[IroncladMarshmallow]

Einstein, I'm curious since you said it's trial and error... Do you have an LC meter? If so, did you measure the capacitance, inductance, and resistance of the oscillatory part of the circuit and calculate the resonant frequency? If it's tuned to a different resonant frequency than what it's oscillating at, it could cause unexpected behavior.

 

[Einstein]

Yes, I have an LC meter. And I did do calculations and compare them to the measured values. They were very close to each other. So I knew what to expect. Or at least I thought I did. This is feedback controlled. So it is always in tune. I thought that was a really cool idea. All my other Tesla coils have to be manually tuned for best operation.

But I didn't know walking close to a Tesla coil changes its resonant frequency. The sensor I made was showing me that on the scope. So basically you could cob together any coil arrangement you want. And the feedback control circuit will adjust to whatever frequency the coil wants to operate at.

Of course matching impedance's does give a better output.

Let me give you an example of something I tried this morning that allows the coil to run in continuous mode. I have a 1 meg resistor jumper wired in parallel with the capacitor and diode on the voltage side of the input. The coil starts up very easily. I removed the jumper wires and soldered the resistor in place. Now the coil wont start at all. I've never come across anything like this before. Common sense tells me to try connecting the resistor with wires the length of my jumper wires. I'm going to try that next. But this is like playing cat and mouse. Only I don't know what the rules are. I should probably right down the things I do that do make it work.

 

[IroncladMarshmallow]

That's really weird. So a temporary, loose contact like a jumper or breadboard will work, but a soldered connection won't. I have no idea what would cause that.

 

[Einstein]

That's really weird. So a temporary, loose contact like a jumper or breadboard will work, but a soldered connection won't. I have no idea what would cause that.

Well, my common sense approach actually worked. This is really weird. I soldered 2 two foot wire lengths to a 1 meg resistor and then soldered that to the PC board. The Tesla coil starts up now.

The same thing happened to another jumper wire I had too. I have a jumper wire loosely connected to a main power junction and the other end to the input switching MosFet for on-off-on-off operation. That would be the IRF510 in the schematic. With that MosFet always on, the Tesla coil runs in continuous mode. So I decided to make a short one instead and soldered it to the power supply rail on the PC board. The Tesla coil would no longer power on with it that way. So I went back to the previous connection method. Now it works again. So it does appear that the length of the jumper wires is a contributing factor to get this thing to work.

I still have the bread board setup. I'm going to pull that out and start looking at the lengths of wire I used to assemble it.

Possibly there is a sequential pattern that all the components need to follow in getting power. That would be a working theory for now just to see if those kind of rules exist for this thing.

 

[Ayasano]

It's not uncommon for the exact specifications of components to have untinended effects on the circuit. I'm reminded of an experiment that used a genetic algorithm to create a FPGA circuit capable of differentiating between the spoken words "stop" and "go" using logic gates. When the circuit was moved to a different FPGA chip, it no longer worked. The genetic algorithm had managed to exploit the specific quirks of the chip, even going so far as to use components that were completely unconnected from the main circuit, which when removed, stopped the whole thing from working.

On the Origin of Circuits • Damn Interesting

 

[Einstein]

I got an update:

I got 3 new breadboards to try out and see if my old breadboards could be contributing to some of the weird stuff going on. Turns out I was right to suspect my old breadboards. Apparently I got some open circuit connections occurring in the old breadboards. So that's why the circuits don't work when I transfer over to PC board. You would think I would suspect that before 5 failures had occurred. But some things you just take for granted. Anyway I got the breadboards off eBay for 6 dollars. Much better than the robber baron price of 30 dollars plus tax at Radio Shack.

With a new breadboard I was able to rapidly prototype a circuit that works both on the breadboard and PC board. So now I have a stable circuit to power my Tesla coil in on-off-on-off mode.

Of course not all the mice have been captured. The coil was outputting a +/-32 volt field. Now all I can get is +/-16 volt field. I did start redoing all the power connections using heavier wire. And making sure the path to power was equally shared. But the wire I was using was speaker wire. So I don't know if having wires running parallel can double an output voltage. But it was part of the equation. Also when I had the +/-32 volt output, I had that weird phenomena where my hand had to be close to the power connection to get the coil running. Now that is no longer required.

Also at present the Tesla coil is operating at an anomaly point. When I move my hand toward the coil while in operation, the output voltage increases up to a point and then decreases. At the turn around point the output electrode of the Tesla coil starts to emit numerous tiny wispy sparks. It looks like a tiny whirlwind while it's happening. So there is more to this field capacitance phenomena than is presently known about. This is something that needs further investigating. But for right now I have stable operation. So I can move ahead and investigate any new phenomena that may occur when I turn on a rotating Tesla coil field.

Here is the schematic that I am presently working with that gives easy start-up with on-off-on-off operation.

 

[Rawknee]

Any updates?

 

[Einstein]

Any updates?

I do have all four Tesla coils up and running. Currently I'm testing out how to configure the four coils together to produce a rotating electric field.

I made two left-hand wound coils, and two right-hand wound coils. I've been monitoring how a left-hand and a right-hand coil interact with each other. At first it looked like they cancel each other out. So I figured that was probably a combination I wouldn't be able to use. But I awoke this morning with a little voice in the back of my mind telling me to take another look. Like I might be missing something. So this morning I started to adjust the distance between the coils. Turns out it looks like there are certain radius's where the coils work exceptionally well. Kind of like the energy shells in an atom. So I'm glad I decided to look closer.

I'm taking this all in at a slow pace. So as not to miss any important observations. It appears so far that these coils are making standing waves. Who would have ever guessed that a Tesla coil produces energy shells, just like inside an atom.

 

[Ayasano]

I do have all four Tesla coils up and running. Currently I'm testing out how to configure the four coils together to produce a rotating electric field.

Is there a specific reason you need Tesla Coils, or are you just trying to make a rotating magnetic field?

If all you want is to make a rotating electromagnetic field, I fear you're over-thinking it a bit. All you need is 2 or more electromagnets running off AC currents that are out of step. (3 coils at 120 degree intervals is quite common) Where the fields intersect, you get a rotating magnetic field. AC motors, invented by Tesla, use that principle, with a centre rotor attached to a permanent magnet, which rotates itself to align with the rotating magnetic field, producing mechanical work. You would just replace the centre rotor in the design with whatever you want to be influenced by the rotating magnetic field.