Plasma generation using coil NO DIODES....
Thanks Shanti....

Circuit is not much fast, but it can work fast with CDI ignition system and proper capacitor and transformer,
I am trying to reduce number of turns of coil to get much better results.....

See Video
http://www.youtube.com/watch?v=eFupa63jY4M

QuoteThanks Shanti....

Your welcome...but why you thank me, didn't do anything...

Well your new circuit is actually a circuit very like to the ones usually currently used for TIG-welding. If you use this setup actually it would be easier just to take a conventional HV generator (electrical gas ignitor, stun gun circuit, ...) and couple it over the HV-transformer instead of the ignition coil. These are quite cheap any readily available...
But I would add something to protect your cap (e.g. varistor or transil diode) for otherwise it can still get some surges as soon as the arc has established.
If you still use the ignition coil, I just would make sure, that the large spark gap isn't too large, so that it's firing voltage is always lower than what the caps can withstand, but on the other hand, the smaller the gap, the smaller the voltage which will go into the transformer...
Here you use the ignition coil to charge the caps. Well I don't know actually how your driving circuits looks like, but if your ignition coil is allowed to ring, efficiency would be far better if you add a HV diode (just a cheap one, doesn't need to withstand much current) between the coil and the caps, so that you make sure, the caps only get charged and not also again discharged by the ignition coil.

You given this link http://www.myelectricengine.com/projects/mpdthruster/ignition/ignition.html

Quoteif you add a HV diode (just a cheap one, doesn't need to withstand much current) between the coil and the caps,

I will try this....

With 12V for LV dosen't make any change in spark (normal spark no plasma with 12V), It needs high voltage it may be because of transformer coil....

Instead of tesla style first part, why not do a simple mosfet connected to 555 timer for high frequency.  This wouldn't cause all that EMI interference.

EDIT:  Would the normal hv spike emitted by the coil be able to pass the air core transformer?  What about if we used iron core?

QuoteInstead of tesla style first part, why not do a simple mosfet connected to 555 timer for high frequency.  This wouldn't cause all that EMI interference.

What exactly do you mean by that? You need HV in this first part circuit. How do you wanna switch the HV (several kv) with a MOSFET? For that you would need a multi FET stage circuit, very delicate to design and also expensive...
But the principle is just to get a HV to the LV circuit. For that you actually wouldn't even need something oscillating, (although it's more efficient like that). A single shot HV on the primary would be enough, so that this single shot gets to the secondary lv side to arc the plug. Sure you could also wire the transformer so that even a LV at the primary would be sufficient to get HV at the secondary, but for this the secondary would need much inductance. And actually ideally we wouldn't wanna have any inductance at all in the secondary circuit for it blocks the flow of the LV.

@cap: As I looked at your circuit again, something right jumped at me 
Your are using seveal different sized caps in series. without any additional circuitry. This is very dangerous. For usually due to the different sizes the caps will be charged to different voltages. So it can very easily happen, that one gets a lot of voltage...Well I just wanted to tell this...
As about the 12V. Well I just made a rough calculation, and the inductance you have in your secondary is so low, it should not really be the reason why the 12V isn't going...Maybe the spark gap is too wide for the 12V to sustain the arc...
How did you connect the 12V, from what source?