3v OU Flashlight

[d3x0r]

In regards to coil turns, when Akula dismantles the coil he appears to remove off fairly equal lengths of wire with each pull. The count for the first winding is 23 lengths and 26 for the second winding.

Edit: I have now tried new software and the probe does appear to be connected.

the inner coil will have a smaller average diameter and be a lower inductance also...  but 23 lengths is more turns on a smaller diameter.... (no help I know)

[TinselKoala]

that's on 2mH?  (the glare on the scale is kinda bad )


Nitpick; the foil winding is the outside winding.... (the other coil is under it... ) also I think maybe the side /direction of that winding matters...


the outside foil should be essentially opposite the wire windings from the connecting point to the other side... the inner side should be essentially a continuation of the winding... so both the outside and inside foils are connected on the same edge relative to their gap (if it's the left side of the gap on the bottom, it's the left side of the gap on the top)


foil - coil connected to foil's - coil - foil

All right, I have remade the coil to conform as best as I can to these specifications.

What high power are we talking about? If anyone can make a one-watt induction furnace I'd like to see it melt something. The very most power I have seen in this circuit is 3 volts at 500 mA DC and at that, the MC34063 chip runs hot. Nothing else gets even perceptibly warm, ever.

Reading is in milliHenry as usual.

[TinselKoala]

And here are a couple of representative screenshots.

I determined last night that the BUZ11A mosfet doesn't really switch properly, even when the strong sinusoid from Pin 13 is present. To see if the transistor or mosfet is actually switching, you need to monitor the Collector or Drain signal, it's the only way to be sure. I say again: the noise bursts are causing the signal to the inverter which results in the transistor switching, not the other way around. There is feedback to be sure, but it does not seem to be the case that the DC-DC converter chip is responding to being pulled down when the mosfet switches... because my mosfet wasn't switching.

However the TIP3055 does switch properly, as you can see from the scopeshot below showing the collector (inverted) trace.

Note that I am now at 1 ms/div horizontally and the sinus oscillations are in the ballpark of the signal on Akula's screen. It's not stable for me, though, and I have some hypotheses about that, such as my lack of a nearby source that is singing at that frequency.

(These are made with the inner, smaller inductance on the transistor side and the larger on the 340963 side.)

ETA: Disregard the numbers in the "measurement" box.... they are using the entire buffer, I forgot to reset it to just use between cursors. So the measurements are including lots of time of unstable waveforms or even non-oscillating blanks. Please just read the traces themselves using the channel voltage setting and timebase. Sorry about that!

ETA2: I am happy to report that the circuit does not respond any more to the 60 Hz mains frequency. So going to the smaller inductances might have been a right guess.

[scratchrobot]

inductance
http://youtu.be/zvVJH-Py39c

[TinselKoala]

inductance
http://youtu.be/zvVJH-Py39c

Yep, that's what ferrite cores do.

And the gap in the core halves fitting together is usually done to prevent the core from saturating, and it can also be used to regulate inductance as you are showing, in a limited range. Vary inductance and you vary resonant frequency, naturally. A mechanical vibration will cause the inductance to change in synch, so there can be feedback between the mechanical and electrical "vibrations" in the transformator.

You must have a lot of wire on that little coil former.