> On Wed, 10 Oct 2012 09:04:17 -0400, "Michael A. Terrell"
> <mike.terr...@earthlink.net> wrote:

> >John Larkin wrote:

> >> On Tue, 09 Oct 2012 18:01:06 -0400, "Michael A. Terrell"
> >> <mike.terr...@earthlink.net> wrote:

> >> >John Larkin wrote:

> >> >> On Tue, 09 Oct 2012 12:35:49 -0400, "Michael A. Terrell"
> >> >> <mike.terr...@earthlink.net> wrote:

> >> >> >John Larkin wrote:

> >> >> >> On Tue, 09 Oct 2012 10:38:17 -0400, "Michael A. Terrell"
> >> >> >> <mike.terr...@earthlink.net> wrote:

> >> >> >> >John Larkin wrote:

> >> >> >> >> Why not put the LED right across the crystal set LC tank? The tank
> >> >> >> >> needs capacitance anyhow. We have some high-efficiency LEDs in the low
> >> >> >> >> 10s of pF. The classic crystal set cap was a 365 pF variable.

> >> >> >> >   That would reduce the 'Q' to an unusable level since a parallel L/C
> >> >> >> >circuit is high impedance.

> >> >> >> The Q would be reduced in proportion to how much power is extracted by
> >> >> >> the LED. Delivering power into the LED is the whole point, so
> >> >> >> naturally that will reduce Q.

> >> >> >> Of course, the impedance match should be optimized for maximum power
> >> >> >> transfer. Crystal sets often used tapped inductors for that reason, so
> >> >> >> thet might be reasonable here.

> >> >> >   And the voltage would be lower at that tap.

> >> >> Depends on what you connect to which tap.

> >> >   Have you ever seen a _working_ crystal radio with the capacitor
> >> >connected to the tap?

> >> No, but we're making a night light, not a crystal radio. And I didn't
> >> say that the cap had to be connected to the tap, but it might make
> >> sense.

> >   Then build it, and show us.

> Not this month; I have a day job. The big hassle would be stringing up
> the antenna.

> There's little doubt, from the math, that it would work, at least
> enough to get some visible light out.