@goat

I appreciate your input, I'm open to all ideas.   When I asked Marco about specifics of his graphite spark gap, but he was sayings its been years and  there are different configurations for low voltage spark gaps.  You'd have to make it small enough that it exceeds the dielectric strength of air as groundloop mentioned.  I have a draft schematic I can draw up , but it's basically just a spark gap on the output stage going to the capacitor in parallel with the LED.

I think the spark gap might work during the higher voltage transients because of the waveform of these joule thief circuits. In between the transients, the voltage would not be sufficient to jump across.
(http://img219.imageshack.us/img219/7654/frequencygb3.jpg)
V/div = 0.2 , T/div = 10us .. but those transient peaks are real high, anywhere from 2V - 18V depending on where in the circuit you attach the probe.  From my joule thief replication using groundloops trifilar circuit.

Also  sorry to be unclear, I dont think the resitance is not an important point at all, I just happened to mention it because I ran some tests on mechanical pencils.  The main reason I'd use graphite is I can calibrate it under a microscope to however many micrometers are necessary, and I have a whole pack of mechanical pencils sitting on my bench.  You could probably do this (calibrate spark gap) too with the NE lamp as well if it was on a fixed surface and could fit under a microscope.  The NE lamp idea is definitely worth investigation.

Again, Thanks to all for the thoughts

Hi,

If you want to use a classic spark, you need to make sure the voltage potential is sufficient to make it over to the other electrode.
Just raise the voltage untill it flashes over, and then lower the input so it stays exactly between the point of spraking/not sparking.
Below is a NE2 bulb attached to a capacitor via two ceramic HV diodes.
The little black stick is a carbon rod for use in a pencil.
When the spark hits the other side (the diodes) of the spark gap the electrons are accellerated before they hit the metal.
This charge is effectively seperated by the diodes and stored into the capacitor.
Be aware that it WILL CHARGE UP TO A LETHAL VOLTAGE!!!
Dont touch it just because you think there isn't that much energy there.
When the capacitor builds up its charge it will get to a threshold of app. 65 volts where it ignites the NE2 neon.
And if the circuit is tuned right it will keep lit depending on the lengths of wire used in the circuit.
If you switch a wire ALWAYS re-tune the circuit.
This you can do by placing another NE2 bulb just before the gap.(one one wire or just wrap one lead around the signal wire)
This bulb will not discharge but it will only shimmer when the circuit is at it's highest potential.
If the voltage is not sufficient, you need to use mechanical spark gap switching.
I have also found an easy way to do this by using two little DC motors pulled out of cd-rom drives.
Magnetize the needles and attach one of the wires to the case of the motor.
Put the two in series and make sure the needles touch each other.
Both motors will run away, one clockwise and one anti-clockwise and there will be a nice mechanical spark in between.

M.

Marco,

In the second photo, the motors appear to be wired in such a way that the HV actually passes through them, with opposite wires connected to the case of the motor and the other ones leading away. Is that correct?

You also say to magnetize the needle, do you mean the rotor shaft?

Thanks.

Well i always see a tiny blue like light between the electrodes.
The NE2's lite up a bit in open loop and they lite up a bit in closed loop.
But they lite up brilliant when the spark is present.

M.

@Feynman,

Something with this circuit is not right. When you press the button then you are grounding the battery through the coil. I have made a drawing that is a little better. The 1K resistor is moved to the ground side of the trigger coil. Now when you press the start button, then you have the 1K to ground AND the coil resistance and inductance. This will not short circuit your battery and the circuit will start to run. Also note that the trigger coil and the power coil must be connected 180 degrees to each other. See the small black dots on the coils.

[EDIT] After the circuit has started to run you release the start button. No need to hardwire.

Hope this helps.

Groundloop.

Here is me replication of amigo's circuit.   We will fix this soon! 

(http://img134.imageshack.us/img134/469/amigo1vy3.jpg)
C'mon... light, little one!!!

(http://img134.imageshack.us/img134/6851/amigo2aa8.jpg)


And here is my replication of groundloop's circuit
(http://img100.imageshack.us/img100/5909/groundloop1mu9.jpg)



I will post scope traces as soon as I can locate an ample supply of beer.  Since it's after midnight, this might require a trip to local bar.    
EDIT:  I just found some Guinness.   

For the spark gap , I am trying to reach a contact at a university lab which has microscope. Suffice to say we can thank JetBlue and the generous budget of Duke University. 

Did anybody yettry toimplement a sparkgap from dissimular
metals like copper and graphite into this circuit to
have a real recharging effect of the battery ?
At these low voltages you have to use probably mikrometer distances.
Or use a higher turn additional winding where you have High Voltage from the transformer
and feed these sparking pulses back to the battery.

It will be then based on the Newman effect and is a direct conversion of graphite ( carbon)
into electricity.
Regards, Stefan.

Marco,

In the second photo, the motors appear to be wired in such a way that the HV actually passes through them, with opposite wires connected to the case of the motor and the other ones leading away. Is that correct?

You also say to magnetize the needle, do you mean the rotor shaft?

Thanks.

Hi Amigo

The mechanical spark gap as shown with the DC motors is not for use in a HV setup.
They can be used in HV setup but then you need to spin them with a battery or something externally.
The motors are wired in series and though their cases in such a manner that when the rotors touch each other, the motors start to spin.
By magnetizing the tips that are touching each other you can keep it running fine but you can also mount the motors with tape or whatever.

You can raise the voltage by using more numbers of turns and redirecting collapse via diodes.
This way you can easy put in like 9Volt and get 200Volt Out as a spike.
M.