Joule Thief

[crowclaw]

May have to include base  limiter resistors to adjust current draw. Ideally circuit components need to be matched, i.e. inductors / resistors.

[Dave45]

Xee2 posted this schematic for a limited self runner, it only charges one side of the cap.
I finished it, havent tested it yet,,,,soon.

[MarkE]

Dave, Neither circuit charges the capacitor from the oscillator portion.  Both charge the capacitor from the battery.  There is no such thing as charging one side of a capacitor anymore than it is possible to stretch one side of an extension spring.  Xee2's circuit draws power from the capacitor through the transformer secondary and the transistor when the transistor is on.  When the transistor turns off, energy stored in the transformer recirculates through the LED.  Your dual circuit just puts two of Xee2's circuits in parallel, where one switches the low side like Xee2's, and the other switches the high side. 

[Dave45]

Hey Mark
I hope your wrong my friend, I see things a little differently.
I'll build it tomorrow.

[MarkE]

Dave, it is basic theory that has been tested countless times.  A test that you can do is to add switches so that you can run with just the NPN, just the PNP, or both the NPN and PNP sides of your device.  You should find that the run down time is faster with the PNP than the NPN due to a lower efficiency.  You should find that with both of them, the run down time is a bit less than half the run down time with the NPN side alone.

If you have an oscilloscope and a low inductance current sense resistor, then you can monitor the capacitor current by placing the current sense resistor in series between the negative side of the capacitor and the est of the circuit.  The scope probe ground clip goes to the low side of the capacitor.  With either just the NPN or PNP side running you will see a saw tooth current drawn from the capacitor.  With both running, because they are not synchronized, you will see two asynchronous sawtooth waveforms that move in time relative to one another.

The modification that you have made to the current flow when the transistor is on is wrong.  Positive convention current flows from the voltage source: battery or capacitor through the coil and through the transistor.  That builds up energy in the magnetic field in and around the transformer.  When the transistor turns off, energy stored in that field causes the voltage across the inductor to flyback until the LED provides a path for the current.  Energy in the stored winding inductance then discharges through the LED.