50kV sparks and plasma from ignition coil.

Input power around 40W this time, single IRF740 switching, square wave 24V, frequency and duty cycle tweaked. The plasma spark is in the last part. Basic circuit here: http://www.niksula.hut.fi/~plahteen/misc/old_projects/ignition/final_stage.gif

Longest sparks here:
http://www.niksula.hut.fi/~plahteen/misc/old_projects/24V_ignition_coil.jpg

In order to get long sparks you need to feed the coil enough power for a single pulse (meaning long enough on-time to build the current). You also need to switch the current off fast at it's peak. That will result in inductive kick which will rise to infinity regardless of input voltage as long as enough current is flowing at the switch off time. However it will be limited by the switching device maximum voltage. You want to let the voltage fly as high as is safe. Secondary voltage is multiplied usually over 100x the primary so 400V kick should be around 40kV at the secondary.

This coil seems to have higher multiplication about 130x.

Some people have had problems with high voltage spikes destroying MOSFETs so if you decide to try and want to be safe you might want to use trazorb such as 1.5KE400A or similar with the capacitor just to be on the safe side. Also use small duty cycle to prevent the tranzorb from heating.

The MOSFET avalanching at around 400V offers a natural clipping limit. As long as that pulse energy is (s)low enough it will survive. The parallel capacitor (s)lows down the kick and helps to protect the MOSFET. Too large capacitor will however result in too large short at the moment of switch-on.

Better circuit here:

http://www.niksula.hut.fi/~plahteen/misc/old_projects/ignition/circuit1.gif

More description of this circuit in my video "Ignition coils can handle quite high power levels".

Single bipolar transistors (such as 2N3055) will not work (well) because they have low current amplification and will act as a resistor rather than a switch. Bipolar transistors aren't voltage amplifiers, they are current amplifiers and will simply multiply (2N3055 maybe 20x) the current they are given (through a resistor). MOSFETs on the other hand have near infinite current gain (with DC) and will act as a switch wasting no power at all.

Some plans with 555 in place: http://www.niksula.hut.fi/~plahteen/misc/old_projects/ignition/circuit4.gif

Yet another (a bit simpler) way of handling the kickback is:
http://www.niksula.hut.fi/~plahteen/misc/old_projects/ignition/circuit6.gif

R and C need to be determined so that the inductive kick will be below the maximum voltage (400V). R is hot.

Category:

Howto & Style

Tags:

• HV
• kilovolt
• spark
• electricity
• ignition
• coil
• high
• voltage

[alex681219]

I use an old audio oscillator, i think generates sinusoidal waves , theres diference in the performance vs. square waves?

[decod31]

@alex681219 In general yes since the spark is formed at the moment of sudden switch off which results in a voltage spike of around 500V which transformed by 100:1 gives you 50kV. Sinusoids don't have that sudden switch off. Not to mention they aren't very good for switching mosfets on and off anyway. If you were to drive the coil directly as a transformer you would need quite high voltage to begin with.

[alex681219]

@decod31 My audio oscillator is a HP200AB model 1950, I think that maybe produce square waves because my ignition coil works fine (8mm) sparks with single toshiba 2n3055...I will try to conect my granpa oscilloscope to see the real wave form.

I have a Motorola mjl21194 transistor, I read that is beter (many bedinis and backEMF collectors use that like ellite transistor)but expensive $ 8 vs $ 2 dollars 2n3055, what you recomend try to fry $8?

[decod31]

@alex681219 Sounds like your audio oscillator might be quite powerful and drive the transistor to saturation and possibly switches it off quickly due to saturation giving you more square wave like waveform. I wouldn't risk the $8, it's unlikely that much better. Audio drive in general is somewhat inefficient, but will work given enough power. Mosfet (IRF740 should be less than $2) with square pulses gives better bang for the same power, but whatever works... ;-)

[southernclassik]

Sounds like little farts.

[decod31]

Hmm, never thought of it that way, probably because I heard live how loud it was, but now that you mentioned it... ;-)

[fdoca]

So, what you think is batter way?

you offer no solution

[lazzer408]

n-ch mosfet gate drive is simple. Pull the gate to source with ~10k and a 1-3ohm resistor in series will help interference. Drive that with a 2 transistor totem-pole. wa-la.

[wildhart7]

one word MOV lol metal oxide varistor , get one!