The Adventures of Skinny Coil

[Fumeaux]

Hi Guys. So this is my first Tesla Coil Thread. I've started a few thread already concerning my coil, but they were never centered around the build as a whole.

This paragraph is a short backstory about me and my coil. You can skip it without missing out on any technical details. So who am I? I was born in the city of Vienna. Went to school and now I study IT. It was when I was around 16 years old when I first came into contact with tesla coil (no direct contact). My Physics teacher comes from Croatia and guess who was also born in Croatia? Yeah right, my teacher. I guess that's why he was always talking about tesla and his inventions. That's when I googled tesla coils and realized the classical tesla coil was quite a simple schematic and I thought I can build one too. And building one I did. First came designing. Smart me thought transformers get higher voltage with more turns therefore I need a lot of turns. Playing with javatc I ran a few designs until I settled on what my current coil is. 3000 Turns and as copperwire is expensive I made the coil on a 7.5cm pvc pipe and ~45cm of turns. The absolute worst part about my design was the winding process. I swear to god it took me 16h handwinding without any pipe holder. I guess not many people can say they rotated a pvc pipe 3000 times. There was another not really bad nor really good aspect of my design. It was always intended to have another pvc pipe (11cm) over my coil itself to protect against sparks damaging the coil. I have to say that is works good against sparks, but 3000turn of 0.15mm copper wire have quite a high resistance which then intern produces quite a lot of heat which gets trapped by my outer pipe. All in all a sub optimal coil, but it is what it is. For now, it works. I've already learned a lot from my project and want to see how far I can push my little coil. So here they begin, the Adventures of Skinny Coil.

Some specs:
-) ~3000turns of 0.15mm dia copper wire
-) 7.5cm dia pvc pipe with winding length of ~45cm
-) ~130kHz (old toroid topload)
-) ~200kHz (new topload)

skinny coil.pdf

The driver:
-) modded Steve Ward 5 (sec feedback, flip-flop, 1000uF on the 12V rail, ucc3732x heatsinks)

The bridge:
-) chinese "irfp460" (blown)
-) irfp450 (blown)
-) fgh40n60smd halfbridge

The spark:
70VAC -> 5-6cm

Current news:
2kVA variac on its way (should arrive friday)

[oneKone]

Can't wait for future posts!

Is you driver interrupted or constant wave? The way you described heating in the secondary seems like your running high on times or CW?

[Fumeaux]

It is running interrupted and personally I don't like to put too much stress on my silicon, which is why i prefere shorter ontimes. I'm more into longer sparks. I noticed the heating running the coil about 10min with long (nearly cw) ontimes. For now this doesn't worry me, but with 3x the voltage there comes 3x the current and therefore 9x the wattage. So 230VAC should be fine too with shorter ontimes, the only thing that bothers me is that the coil gets warmer than the heatsink.

[Fumeaux]

So my Variac has arrived. Everything work alright, there is just one Problem I have noticed. Whenever I "draw" an arc the frequency increases to 400kHz which is kinda odd. This results in a different kind of arc which is much narrower and blue instead of purple/white, but about the same length. Strangly there are also racing arcs, just when I draw the arc. I think this is what killed the first two of my IGBTs. As long as I don't touch the arcs everything is fine.

The parameters of the fail:
260VAC input, longest ontime and quite short offtime. Strike to a piece of metal in my hand.
Both IGBTs have now a resitance of <1Ohm, so I guess they are just dead, even without visible damage. My input fuse (5A) also died
The heatsink was cold, the plastic of the IGBTs slightly warm, the diodes (1560) cold.

Some information about the driver:
It is loneoceans sstc3 v1.0 driver without the UVL and with some MUR 1560 diodes to "protect" against voltagespikes. There is close to 0 deadtime and I'm thinking about delaying one of the UCC's with two of the unused hex schmitt triggers, to get about 25ns of delay. Any ideas on this?

Luckily I bought two sets for my halfbridge and it is running again.

So this is the best result I ever got from any tesla coil and therefore I'm quite happy.

[Bambinz]

Very nice
I think that you can obtain better spark with your tesla.

Have you calculated the topload to obtain quarter wave frequency?

[Mads Barnkob]

Sound like you have some strong harmonics since your driver can catch on to 400 kHz for a 200 kHz system, maybe you have to add some kind of low pass filter to your feedback CT, but that should not be necessary.

How about excessive noise making the driver behave odd? Is it inside a grounded metal case?

I do agree that for 260VAC you should see at least 20-25 cm sparks with a SSTC.

You also mention racing sparks, that usually a sign of too high coupling or very bad tuning.

[Fumeaux]

I guess a lowpassfilter might work.

It wasn't in a case, but I figured as it was running fine with sparks to open air, "ground" ark wouldn't have a different effect on the driver. Now I added a ATX psu shroud over it, to give it atleast some protection. I couldn't test it yet, as my variac constantly trips my circuit breaker, which I have to fix first.



The racing arks weren't like the normal racing arc you see in tc's. At lower voltage (120V) I started to get a lot of corona from the upper primarycoil connector towards the topload, when having a "ground" ark. At higher voltage the corona would go back to the secondary coil.

I think this is not due to a high coupling, as it would have been a constant problem. Actually I think I should increase the coupling, as I currently have quite a low one. It must be around 0,3. So it must have something to do with the "ground" arks and them detuning the driver. I hope the shielding helps. If not, I will throw a lowpass filter together.