MMC capacitors exploding on my tesla coil, I

[Crazy Coiler]

MMC capacitors exploding on my tesla coil, I need help!

Im in a tight spot. I just recently built a decent sized tesla coil. Everything works flawlessly except that my capacitors are melting. They don't fail but they emit flames or they arc, (its hard to tell). Then after inspecting the capacitor bank, you can see that the capacitor has melted through. It still works but soon the capacitor catches completely on fire with flames licking the other capacitors. The curious part of this entire phenomenon is that its the same capacitor in the same spot. No matter how many times I replace it, it first glows, melts, and then burns. However, for whatever reason, instead of melting, the resistor just disintegrated instead. I have also noticed that when running my coil at night, 2 or 3 other capacitors have this bright ominous blue glow on the ends. The capacitors are rated for 6.5 kv DC which is more than enough. I have 7 rows in parallel with 5 capacitors in series. My NST's are each rated for 15 kv @ 60 mah but wired in parallel for 120 mah. I have replaced about 8 capacitors so far. 5 came from the same spot while others just ocurred from the spark gap being too large.

 

[0xDEADBEEF]

[...] The capacitors are rated for 6.5 kv DC which is more than enough. I have 7 rows in parallel with 5 capacitors in series.[...]

I haven't build Tesla coils yet, so I'll try stating the obvious.

I don't know the exact setup but as far as I remember for a tesla Coil you are discharging a capacitor into a coil which should cause things to swing as an RCL oscillator. 6.5kV DC supposedly means it will hold 6.5kV when charged, but it does not tell you anything about the power rating or the current. If you discharge the capacitor into a very small coil, with a low self inductance, the current will be enormous, if the wiring inside the capacitor has too much resistivity, then it will heat up, typically getting even more resistive, causing small hot spots.
I am quite sure, I can burn any capacitor with 6.5kV DC rating with a few thousand Volts signal if you allow me to raise the frequency high enough that the impedance of the capacitor comes close to 0. Try putting a resistor in series with you capacitors, or put more turns on your coil.

The blue light sounds like a corona discharge (http://en.wikipedia.org/wiki/Corona_discharge) try to avoid sharp points.

 

[Crazy Coiler]

Sadly the caps don't have any sort of power rating or current. My prognosis is that the caps are cheap. As for discharging into a small coil, I'm running a 6" coil with about 1500 turns on the secondary and 10 on the primary, it's a decently large coil. I'm not 100% sure what your saying with the low self inductance. You can read my mind. I end up with small holes in my capacitor that then emits melted oil, etc. But my big question is why is it happening on ONE capacitor. I replace it, plug the coil in, and ten seconds later, that same capacitor fails.
I know what corona discharge is, but coming from the capacitor? I mean my solder joints aren't the best so maybe the leads that are sticking out are producing the corona? Or are the caps popping open and letting air into the dielectric area? Luckily it is only on one capacitor.

But this is interesting, I just redid that one string of capacitors, and plugged it in. Instead of the capacitor failing, the resistor attached to it melted. They are just bleed resistors.
So your saying I should put a resistor in series? This may seem to be a stupid question but where in series?

 

[0xDEADBEEF]

1) Where to put the resistor:
You could put a low Ohm high current rating resistor in front of the primary coil, to make the capacitors discharge more slowly. This is tricky though, the current is very high, and the skin effect seems to produce unexpected resistance effects according to my book. Maybe try point 2) first.

2) Did you check out http://deepfriedneon.com/tesla_f_mmc.html they have some nice info on capacitor arrays. Where they state that their capacitors AC rating is less then a third of the DC rating. I don't really know much about these ratings, as I told you, I am sure that you can burn pretty much any capacitor if your frequency is high enough, because then they act like shorts, but I would guess your array is still a bit too weak, and that the rating you should work with is somewhere in between AC and DC. From their info I take it, that your voltage might peak at 30kV so your 32kVDC setup doesn't allow for much performance loss.

3) Why it is always in the same spot: This is fairly typical for these run away resistance effects. I would guess in the affected string the capacitors are exceptionally good, having a low internal resistance and high capacity or one might be shorted, so the current in that string is the highest, and the capacitor you put in is the weak link, which will short and show a runaway behavior. It is strange that the bleed resistor should melt, unless the capacitor had bad contact.

 

[Crazy Coiler]

Apparently my message didn't post yesterday.

I have a spark gap which controls how much the capacitors discharge. It uses air as a resistor and is adjustable. Even at the smallest "setting" I am blowing capacitors. I actually used deepfriedneon for my initial layout of my capacitors. I took the data they gave me and that's what I used to buy my caps.

In your second point, are you suggesting I add more caps in series to up the overall rating and then add another corresponding string of caps to keep my capacitance the same? If the rating is actually 1/3 what the label says, then my bank is only rated for about 10kv. Which would be a huge problem. Would you have any recommendations for adequate caps? The only ones I can find are rated DC and none of them are rated anywhere near to the ones I have now.

As for the third point, are you saying I should replace the entire affected string? I was focused on the one capacitor and kept replacing it, I didn't think to replace the whole string. Or is there any way to tell if the cap is faulty? Keep in mind I now have only two spares left to use so I have to either order more or completely redo my cap bank...

 

[Bob S]

If you analyze the LC circuit (tesla coil primary and your capacitors), although you are charging the capacitors with a dc voltage, as soon as you discharge them into the coil, you have an ac current and voltage from an LC resonant circuit. We used to use either oil-filled or doorknob capacitors.

 

[Phrak]

How can you get coronal discharge across the capacitor terminals of a capacitor designed for X volts, when X volts are placed across the terminals? Is your humidity high? Are you using dialectric cups (plastic booties) over the capacitor connections?

 

[0xDEADBEEF]

[...]
In your second point, are you suggesting I add more caps in series to up the overall rating and then add another corresponding string of caps to keep my capacitance the same?

Exactly. It is easy high school physics, to calculate the number of resistors per line and the number of lines you need, once you know which rating you should use.

[...]
As for the third point, are you saying I should replace the entire affected string?

If the rating is really much to low the electricity will simply look for the second weakest link... but you could give it a shot.

I was focused on the one capacitor and kept replacing it, I didn't think to replace the whole string. Or is there any way to tell if the cap is faulty? [...]

I don't know what kind of equipment you have. You should at least have a multimeter. So you could check for shorts over the caps. Some multimeters can also measure capacitance. Oh and by the way I hope you are not using elcos