Tesla Coil Build Thread

[Seoul_lasers]

It's the peak current that you need to drive all those windings. You need large capacitors and a MOSFET driver.
A very small 24V DC powered MOSFET TC has been built but the circuit is a bit complex. Skori Weblapja - Mini Tesla tekercs


Video:

 

[Sigurthr]

That is REALLY excellent performance for 24V supply, but I think the pulsed nature is to thank for it. What is truly amazing is the low coupling between the coils! I've never seen Solid state non-DR coils use such low coupling.

 

[Seoul_lasers]

I think this little coil is a record breaker for smallest coil with
the highest output/ streamer length. If memory serves me right.
Tempted to build it!

 

[Hiemal]

It's the peak current that you need to drive all those windings. You need large capacitors and a MOSFET driver.
A very small 24V DC powered MOSFET TC has been built but the circuit is a bit complex. Skori Weblapja - Mini Tesla tekercs


Video:

What may be a neat idea is to take the same coil, and use a boost converter to bump up the 24 volts to something higher to increase the power. Though, the boost converter would need to be pretty chunky to work properly!

 

[Atomicrox]

That looks great but I think it's a bit too complicated for my skills. I was thinking something much simpler, probably a single MOSFET driver.

Also found this one:
The Mini SSTC

 

[Sigurthr]

Ya, the Mini SSTC (half bridge w/ UCC chips and a schmitt inverter) is a rock solid design that I use for all of my solid state TCs these days. It is cost effective and easy enough to build with no real pitfalls.

The single ended (one mosfet) designs I've seen have all given me tons of trouble, invariably all eating numerous mosfets. Basically it's like trying to use a flyback driver to power a tesla coil.... the flyback voltage either destroys the mosfet or you isolate the fet and reduce output severely (as well as generate tons of heat in the isolation diodes).

I actually have a revised version of Steve Ward's SSTC 5 if you would like to use it. It is my class D amplifier design basically, you just use the tesla coil resonator as the primary/secondary and ignore the output filter on the secondary. It has enhancements such as single voltage rail logic (12v only), gate protection, full body diode isolation, and built in audio modulation. I can draw up a schematic if you like. I can post my class d amp schematic but for TC use it requires you route the sg3525 subcircuit output to Pin 3 of the UCC drivers instead of a voltage divider leading in to Pin 1 of the schmitt trigger inverter, then just connect an antenna to Pin 1 and use the clamping diode setup shown in Steve Ward's SSTCs from pin 1 to the logic supply rail and to ground.

 

[Atomicrox]

I hadn't read the rest of the thread before my previous post, now I've read most of it

I'd love to see your improvements but I'm afraid I won't be able to buy all the components to built it here, specially the GDT toroid, both gate drivers and those "special" diodes.

I'm *really* bent on building a TC but my options seem pretty limited:
-Use my 555 flyback to power an SGTC.
-Build the mini SSTC, I'll have to ask someone here to buy the hard to find components on the USA and ship them to me.
-Try to make some sort of single MOSFET driver and end up frying lots of MOSFETs.

:/

 

[Sigurthr]

You can use a beefy flyback to make a single SGTC, it works rather well and I've done it many times. The hard part is two-fold:

1) you need a primary winding and coilform which will insulate against the short-circuit voltage that your flyback transformer outputs (depending on drive method and transformer, anywhere between 1kV and maybe 30kV)

2) for optimal results you need to match the impedance of the primary LC circuit to the characteristic impedance of your flyback transformer, which is unknown.

The good part is that it is isolated from mains, runs from low voltage input (battery capable), and is severely limited in current. So in the end it will be a rather safe coil, but output would be tiny by most standards, still a functional real TC though.

As for sourcing parts, I can't help there, but if you do decide to do that I highly suggest you use a modified (refined/improved) design from the one Steve Ward posted. It is an excellent design, but it is not "finished" in my opinion. When I get a chance I'll draw up my revised version.

 

[Atomicrox]

1) I can probably solve that with another PVC tube between primary and secondary.

2) Yeah, that I don't have the slightest clue :/

Since I have all the stuff here except for the tube and the magnet wire I might give it a shot, probably using a leyden jar capacitor...


I always forget about ebay.. a quick search found most of the hard to find components there, except for the gate drivers. Is there anything I can substitute? Perhaps a BJT totem pole or TC4420/29 (those are on ebay)?
I'll probably give the SSTC a shot if I can find substitutes for the gate drivers. Looking forward to seeing your modifications!

 

[Sigurthr]

Re: 1) you don't just have to insulate the primary side voltage from the secondary, but from itself too! If you supply the primary coil with 33,000V DC you need to have 33,000V/n (where "n" = number of turns on the primary coil) of insulation between EACH TURN of the primary coil. If you power a primary coil with AC you need to have (1.414 x Vinput)/n insulation between each turn.

Re. 2) I've used TC4420 (but not 29) before with good results. The problem is that this chip doesn't have an enable pin so you can't do audio modulation via Logic side PWM like you can with the UCC chips.

Eastern Voltage Research - Electronic Kits, Tesla Coil Kits, LED and LED Driver Kits, Home of the DRSSTC sells the UCC chips internationally -AT COST- by the way, which is an exceptional deal. I've had business with the owner there many times and he is quite excellent as well. Just go to the "semiconductors" page on the site.

 

[Atomicrox]

1) Indeed! A loosely coupled primary with ~1cm between turns should do it but gotta be extra careful..

2) 14.50 USD for both + shipping.. Not cheap but I can live with that price, thanks!
(other sources would be appreciated, in case anyone knows)

Have you tried audio input on yours? I'd just use pin 5 of that interrupter 555, right?


How do you guys choose your MOSFETs? I've seen HV projects using IRFP250, 260, 540, 450 and 460. Is there a table with their characteristics (besides individual datasheets)?

 

[Sigurthr]

Placing space between the turns of a primary coil reduces the inductance which will cause the coil to draw more current and if used in a non-autotuning design will cause a loss of output from detuning the coil. In a standard SSTC with antenna feedback it shouldn't affect tuning, but it will require additional turns on the primary. Solid state coils require a very tightly coupled primary to secondary, by the way. Only high power DRSSTCs and spark gap coils function well with loose coupling.

There isn't really a chart as there are thousands of options. Generally you pick a mosfet by looking for one with a Vds MAX (blocking voltage maximum) >1.25 x your supply voltage and then look for as low of a RdsON (on resistance) as possible. You need to choose one which can handle the peak current safely, which is Id MAX (drain current maximum). Again aim for at least 25% overhead. If you are going over about 500KHz you need to look at the gate capacitance (gate charge) and switching speeds (turn on delay, rise time, turn off delay, fall time) which gets a bit more complex. Smaller numbers for gate capacitance and switching delay & rise/fall times are better.

Also, there is more to a mosfet than its datasheet. IRFP460's have been known to fail for no apparent reason in TC environments. The flipside is they are abundant and cheap. IRFP260s are much hardier and if you are only using 120V mains or less, will work fine and replace the 460s. I'm partial to fairchild's FDL100N50F mosfets as they are real mosters at 500V 100A TO-264 and very robust. I use them in my Harmonic-Output-SSTC, which operates are more than one frequency simultaneously. This punishment would blow 99% of other mosfets, but they run just fine even on semi-small heatsinks. The flipside is they are very expensive at $25 each.

Ohh, and those UCC gate drive chips cost me $4/ea + $8shipping here in the US, so keep that in mind. If you can afford it I recommend ordering at least 4 total, they are VERY static sensitive and require decoupling caps right at them when in a circuit.

I've never tried the 555's pin 5 for audio, but then again I don't use the 555 part of Ward's circuit as it is very poorly designed. It alters pulse width and frequency but the two controls do not operate independantly. Changing one will change the other and the entire range is pretty useless aside from making a loud noise. His 555 subcircuit is an interrupter which operates in the audio frequency range, I think between 200Hz and 1200Hz. For audio modulation via interruption (be it enable pins or a standard interrupter) you need the interuption frequency to be at least 20KHz. The higher the better. An interupter for streamer growth needs to operate between 50Hz and 300Hz as this is the optimum range for growth.

 

[Atomicrox]

Any videos of your Harmonic-Output-SSTC? I think I've never seen one of those!

Seriously? I thought those things were very cheap on in the US. Guess that site's price isn't bad after all!
Any other components you recommend having spares (besides the MOSFETs, that is)?

I think I'd better wait till I have more spare money before I begin this project, it's looking far more expensive than I thought


As for the 555 I don't really get that design of his. The standard astable layout is more understandable and easier to get the frequency/duty cycle you want...

 

[Sigurthr]

I think in his original designs he used something other than ceramic disc, polypropylene foil, and electrolytic caps. Those three I mentioned are the only types safe for TC use, all others will explode violently (I've had it happen!). Every IC needs an electrolytic decoupling cap directly across the power input leads of the IC, as physically close to the IC as possible. Non-polar electrolytics are best, but polar are OK providing no outrageous errors are present in the overall system.

Other than the mosfets and gate drive chips you should be fine with a minimalist parts list. I've never had the schmitt trigger fail, and all the rest is pretty hardy stuff.

Yes, his 555 interrupter circuit there is odd. It is a PWM design used for LEDs and such adapted for use as an interrupter... not ideal. Steve Ward's innumerable contributions and advancements over the years well make up for his errors. Hell, one of his first incarnations had like six inverting buffers stringed together for literally NO reason at all.

 

[Atomicrox]

One more safety question, if you don't mind: TCs and other HV stuff usually generate UV. Do you know if it's just UVA/UVB or if UVC is also generated? I googled but couldn't find anything that specific.


Back on his design:
If I decide to build it I'm thinking the best approach is to leave the interrupter on one of those small breadboards so I can change resistors/cap to get a huge frequency range (small with no audio or high with audio). Either that or I'll use a switch with various caps for the 555. Or perhaps it'd be best to use a PWM chip instead?

 

[Hiemal]

Every IC needs an electrolytic decoupling cap directly across the power input leads of the IC, as physically close to the IC as possible. Non-polar electrolytics are best, but polar are OK providing no outrageous errors are present in the overall system.

Actually, I recommend against using JUST an electrolytic. Electrolytics typically have high-ish ESR, and ESR to a cap basically limits how much pulse current they can shove out, and since that's basically what a decoupling cap is there for, that's an issue!

If you want to decouple something, a 1 uF ceramic cap is usually enough (hell even 0.1 uF works!) super close to the supply pins, and you should be good to go.