Joule Thief

[TinselKoala]

Here's my "current" circuit, with the mods added to the basic LaserSaber crossover v.2 circuit. The 1.0 uF cap is to demonstrate the "charging" effect, and also in combo with the resistor across the diode it will produce a neat "flashing" mode.
With the MPSA18 the circuit is very sensitive and I've blown a couple of LEDs by having too much voltage (like over 2 volts !!) on the capacitor when switching modes. But the circuit can run in GL mode with 12 volts on the cap and it takes a long time to run down. With the MPSA18, the light glows in GL mode even without touching, depending on wiring and component layout, but touching brings it to full "GhostLight" brightness.

The lowest current draw I've seen with still visible light has been around under 50 microAmps. This is not yet to the 1 uA level but it is astounding to me anyway.

Note: use a center-off switch for the "mode" switch at the Q1 base; the MPSA18 doesn't need the resistor to start at higher voltages but might at low voltages.

[TinselKoala]

This is very interesting.

I've failed a couple of 2n2222s by having the "mode" switch at GL mode or off, charging the cap to 12 V and then switching to JT mode. With the LED, it blows the LED by shorting it. With no load, the 2n2222 fails, I think. But this latter failure.... is really really weird.

With a normal 2n2222 the circuit behaves "normally". But with the failed 2n2222 and 1 to 2 volts on the cap, or supplied by the battery.... the light comes on and none of the switches do anything! And of course the GL mode doesn't work.

I'll demonstrate this weirdness in a video shortly. Need a cup of coffee. At first I thought that the transistor was simply shorted, and allowing the 2n3055 to act as a single stage JT. But I removed the 2n2222 from its socket and can't duplicate the behaviour with jumpers shorting the transistor pins.

[Groundloop]

Hi,

I have tried to replicate the two battery charger variant of the joule thief.
Could not make it run, then after 4 trial I gave up. Can somebody help me with
this, I would like to make sure I-m doing right the  toroid at least.
Which of the four variants is the good one?

best regards,
sarmasio

Sarmasio,

The circuit drawing (that I made) is correct and the switch is wired correct.
Now, take the red and orange marked wires (not the coil to the led) and twist them together.
Make 10 turns on the toroid. Solder the wires together at  at the opposite twisted ends. This is the
top wire point. Connect the two other free ends to the resistor and to the collector of the transistor.
Make a single wire 20 turns for the LED output. Put a 10K variable resistor in series with the
1K resistor. Put a 1nF ceramic capacitor over the endpoints of the variable resistor and 1K resistor.
Now you can tune the oscillator frequency for best performance. Connect the batteries via a switch
as shown in the drawing. Adjust 10K variable resistor for best light in LED.

GL.

[Groundloop]

This is very interesting.

I've failed a couple of 2n2222s by having the "mode" switch at GL mode or off, charging the cap to 12 V and then switching to JT mode. With the LED, it blows the LED by shorting it. With no load, the 2n2222 fails, I think. But this latter failure.... is really really weird.

With a normal 2n2222 the circuit behaves "normally". But with the failed 2n2222 and 1 to 2 volts on the cap, or supplied by the battery.... the light comes on and none of the switches do anything! And of course the GL mode doesn't work.

I'll demonstrate this weirdness in a video shortly. Need a cup of coffee. At first I thought that the transistor was simply shorted, and allowing the 2n3055 to act as a single stage JT. But I removed the 2n2222 from its socket and can't duplicate the behaviour with jumpers shorting the transistor pins.

TK,

Have you measured (Ohm) the failed 2n2222 to see what failed?

GL.

[gadgetmall]

Nice Going TK !! . Stay away from 12 volts to save your transistors . Now you are almost there if you can fire them on .7 you can try the ground and antenna tesla power .