Tesla Coil Build Thread

[Polonium210]

Sig, I think the heatsink problem has to do with the impedance matching as said above. To fix that issue I would suggest a smaller secondary. 3x7inch wind length of #30awg.
I have found with SSTCs the larger the secondary the more current will be drawn from the bridge with less efficient power transfer.
You can push more power into a smaller secondary more efficiently with better output.
That most likely will solve heating issues.
Also, try and make all your connections from your bridge to your primary as short as possible.

 

[Sigurthr]

I think you're right about the impedance mismatch of the large secondary however a smaller secondary means a higher f0 which means the fets will spend more time in the linear range.

Anyway, my taxes are in so I'd like one of your board sets now please! PM incoming!


@Atomicrox I can't remember off hand but it is the same old formula for Vp-p ripple. You just pick a ripple value which meets your application and put in current, frequency, and you work out the capacitance needed. If I didn't want audio modulation or silent CW I wouldn't need such huge capacitances.

 

[Atomicrox]

Maybe you could wire the multiplier with a switch to use it only on pulsed mode. That's what I plan to do on mine, anyways.

BTW just got a nice pvc tube form, both end caps and a "swimming tube" to make the toroid. Will attempt to either start winding tomorrow or apply some varnish.

Do you guys use varnish before winding? Just on the coil area or do you seal the whole form?

 

[Sigurthr]

I sanded down my pvc with 400grit then used kapton equivalent tape to hold the first few turns. Wound it and then sprayed the hell out of it once it was done with varnish.

 

[Atomicrox]

Here's my secondary:

3"x7"
30AWG
9.6mH

And here's the pool toy that will become my toroid:

2.5"x8"

 

[Sigurthr]

Be careful how you route and terminate the HV end of the secondary over the end cap. I had massive issues with corona and carbon tracking (and the subsequent flames) because over time the enamel on the magnet wire dielectrically broke down where in proximity to the endcap. I eventually discovered that I could "shield" the HV end with thin copper tubing which produced no corona against the plastic.

Looking good though, I can't wait to see it up and running.


I am really itching to build a very tiny SGTC. I still have my old MMC and NST in the basement collecting dust. I just don't really have the time to do it, nor do I feel like winding another secondary, lmao.

 

[Fiddy]

I am really itching to build a very tiny SGTC. I still have my old MMC and NST in the basement collecting dust. I just don't really have the time to do it, nor do I feel like winding another secondary, lmao.

Me too! never built a SGTC, let alone a small one, i found some smallish 240V -> 9kV ignition transformers at work that may be good for them.

 

[Sigurthr]

SGTCs are pretty fun but they are a nightmare to tune and too loud to enjoy for more than 30sec at a time. The main difference between a great running sgtc and a poorly running one is the spark gap. Quenching is key and cannot be stressed enough. Even at low-ish power the primary gap's arc can fail to extinguish with even several psi of airflow trying to break it. To solve this high end coils use rotary or multi-segment gaps, but this adds a tremendous amount of complexity to what is a very simple design. While many successfully use multi-segment gaps to get around the lack of quenching, doing so dramatically increases spurious system losses.

OBITs are cool but I hear they die easily if RF gets back in to them.

 

[Atomicrox]

Thanks for the tips, I still haven't decided exactly how I'm going to attach the toroid...
Probably a plastic bolt thru the end cap providade I can find one. Then I'll screw the wire directly to the toroid.
Else I'll use a regular screw and attach the wire to the screw like most coilers do. I'm trying to avoid doing that to both ends because of internal sparks.

I'll add more pictures as I build it but it'll take quite some time, with my dayjob I only get time for this at weekends :/

My original plan was building a dual-MOT SGTC but I'm glad I went for the SSTC, I really lack the tools to do anything that requires mechanical labor.

 

[Sigurthr]

Oh so you are going the solid state route, you can just bolt on the topload then provided you use a round top bolt. There won't be any internal arcing; arcs from solid state non-DR coils are much less aggressive. You might still have the corona issue I mentioned with the HV end, but the bolt will be fine (that is how mine is attached).

 

[Atomicrox]

How did you attach the bottom end to the ground cable?

 

[Sigurthr]

I used a bolt and a thumbscrew nut, some washers, and another nut. Remember you need to strip the enamel off each end of the secondary wire. Your ground connection should be as thick cable as possible, strapping is better, as this is a high current low voltage connection.

 

[Atomicrox]

OK! I'll try with a tin foil counterpoise before actually grounding it, though.

 

[Sigurthr]

Hehe, be careful if you do the counterpoise test... when I did it with my "baby" 500W coil I set the table it was on aflame. The edges of the foil can have a high voltage gradient and it doesn't take much corona above 100W to ignite/carbon track wood or plastic. Granted, on that coil the primary was adjusted relatively far up the length of the secondary for high coupling for best output which resulted in an altered voltage distribution across the secondary compared to what is "normal".



I finally got some time today to work on further testing of my audio modulation for my big coil (which is now a full 2kW CW). Using the enable pins of the UCC chips is completely out the window. Driving the enable pins with anything over a couple kilocycles drops output power by an immense factor, something like 97% at best. At f0 on the enable pins I was able to get up to around 7% power with full input voltage. So, this rules out the "just not at the right pwn switching frequency" theory.

That being said I did find I can direct drive the coil without feedback as long as I am very close to f0. So now I just need to make an audio pwm circuit that can handle 166.6KHz. I was using a SG3525 chip initially which can only do up to 100KHz. I then rigged up an astable 555 on breadboard to 165KHz and fed that in to the antenna, and it ran the coil fine, albeit a bit noisily probably from some mains leakage or just not being right on f0, but the coil still operated fine. I'm pretty sure I can just ac-pass pump low level audio in to the pin 5 of a 555 to get audio pwm out, but I've never tried it. It will have to wait until I have more time to build that and scope it out. Then it will be a matter of hooking it all up again and testing it out on the coil.

I'm really not sure why I can't audio modulate via my enable pins when I know others have done so with no issue, but at least I'm finding a work-around. If only I had a car audio amp and a 120/12V transformer which could handle 17Amps on the 120v side I would just AM modulate the DC bus and get 4kW PEP output on modulation peaks. That would be one MEAN LOUD TC flame.

 

[Atomicrox]

LOL, I'll try it on a cemented ground then

Your problem with the enable pin is strange indeed, I've seen a lot of people using it for that... in fact that's what I intend(ed) to do!

As you said I was gonna feed 555's output to the enable pin of the UCCs (running on antenna or secondary base CT feedback, around 430kHz), with a base frequency around 20-40kHz and amplified audio fed to pin 5.
That does work quite well on flybacks but they're not as resonance critical as TCs.

 

[Sigurthr]

Yeah I really expected it to work and was quite disappointed and shocked when it did not. I'm still not sure what exactly is going on and the folks at 4HV wouldn't touch it with a 10ft chicken stick.

If the modulation frequency on the enable pin rises above about 2kHz the coil's output drops off exponentially with frequency until within 2kHz of f0 where it starts to increase again. Right on f0 as close as a 555 can get I get about 90% output of normal, which makes sense as the phase difference between f0 feedback and modulation will cause output nulls in the form of beating.

3/17/13
Update: I was able to get some time in with my big coil and play around with a basic 555 interupter again, first time in months. I did a nice frequency response sweep from 10Hz to 1.5kHz on the enable pin. Longest arcs appear at ~90Hz and ~780Hz. At 90Hz the streamers are almost plant stem shaped with fluted thick sections every centimeter or so terminating in sword like sparks. At 780Hz a classic spider web / "banjo" effect is seen. Let me tell you... 1900W of 780Hz requires hearing protection. My ears are still ringing even with using ear plugs.

I also managed to track down the source of the noise when the variac is cranked up over about 80Vac; it is the DC bus "bottoming out" causing stuttering effects. 1800uF simply isn't enough for this bad boy. Good thing I'm using full wave rectification, I'd hate to see it on half wave. What I wouldn't do for a nice 50,000uF 400V cap and some high power diodes for a doubler circuit (I'm drawing 17.86A on the 110V line).