Tim Williams wrote:
Inspired by, hmm, what the hell compelled me to do this anyway? -- I made
one of those 1.5V to white LED inverter dealies. I found some
exceptionally small RF toroids and wanted to see how it would go. Having
not done this before, I started with larger toroids, which went very
nicely. This small piece has (understandably) little inductance, so it
took three times as many turns as the earlier coils, and still the
operating frequency ended up at a whopping 5MHz. The LEDs still light up
just fine. They don't seem to mind what the frequency is.

Toroids work great for that. For fun, try it with an air
wound coil and red led. AIRC the one I made that way used
20t of #22 insulated, wrapped on a 1/4" bolt, then allowed
to expand and the bolt removed.


1.5 ---+------+---------+---------+
| | | |
[4.7K] [100uF] [100R] |
| | | |
| | [LED] |
| | | |
| | /c |
+------+-------| BC337 [coil]
\e |
| |
Gnd -------------------+----------+


Ed


Circuit is (in order): a 0.1uF ceramic cap (the yellow box), a 2N4401 wired
as blocking oscillator (though the blockingness at this frequency is
debatable!), the coil (15T each of around 26AWG, one in series with the
collector and LEDs, and the other in series with the base), and two white
LEDs, claimed as 1100 mcd on the RadioShank bubble package they came in.

Output seems to be limited by voltage and peak saturation current, which in
all cases peaked out at 600-800mA. Collector saturation seems to be in the
0.3V range, which is probably in part due to the large peak base current
possible here. I'm not counting efficiency at the moment, so what the hell
do I care. ;-)

Tim

--
Deep Fryer: A very philosophical monk.
Website @ http://webpages.charter.net/dawill/tmoranwms