Joule Thief

[MarkE]

It is a reasonable guess that the "high voltage" is something he measures when the LEDs are not connected.  It is also a reasonable guess that it is less than 50V peak, which is certainly much greater than the three volts required to run a white LED.

[TinselKoala]

Not necessarily. It all depends. Here's a no-load scopeshot from my HVJT that runs 6 NE-2s in series on one AAA battery.

[MarkE]

That's only because you used a transistor that does not breakdown before the neon bulbs fire.  A 2N3904 or 2N2222A as most would use both breakdown at 40V (some 2N2222A's will do 50V).  I suspect that your 200V/div is really 20V/div.  It should only take about 65V-90V to fire an NE-2.  At 20V/div your pulses would be about 84V, and ~2us/260us * 84V = 0.64V*T_INTERVAL which is within the ball park of a AAA battery that is down to 0.8V at 80% efficiency.

[TinselKoala]

Nope, the 200v/div is the actual true value. I just checked it again with a different oscilloscope and a 100x attenuating probe. No load, the AAA battery reads 1.32 V and with the circuit operating it reads 0.96 V, and the spikes go to about 750V. (With a 10x probe and the channel set to 50v per div the spikes go off the screen at the top.)

And I just put in a fresh battery, circuit off Vbatt 1.53 V, circuit on Vbatt 1.15 V, and with all six neons in place and glowing brightly the spikes top out at over 800 V.


ETA: I just realized I didn't have the "CAL" knob turned all the way CW on the channel. A remeasurement with the CAL knob set correctly shows the spikes going to just under 900 V.

[MarkE]

Nope, the 200v/div is the actual true value. I just checked it again with a different oscilloscope and a 100x attenuating probe. No load, the AAA battery reads 1.32 V and with the circuit operating it reads 0.96 V, and the spikes go to about 750V. (With a 10x probe and the channel set to 50v per div the spikes go off the screen at the top.)

And I just put in a fresh battery, circuit off Vbatt 1.53 V, circuit on Vbatt 1.15 V, and with all six neons in place and glowing brightly the spikes top out at over 800 V.

Then I have three questions:  Is the scope waveform taken from the collector to the emitter of the 2N3055? Are the NE-2's on the secondary side of a step-up transformer?  Are the NE-2's wired in series or parallel?  If the circuit uses a step-up transformer, and the scope waveform is taken on the secondary side of that transformer then the V*T balance estimates will be off by the turns ratio.