Chronology Current Month Current Thread Current Date
[Year List] [Month List (current year)] [Date Index] [Thread Index] [Thread Prev] [Thread Next] [Date Prev] [Date Next]

Re: LR demo circuit



Carl E. Mungan wrote:
timescales on the order of seconds ... L/R is at least
3 orders of magnitude away from the seconds timescale and it seems it
would take a pretty unusual coil to bridge this gap.

Maybe not as unusual as you think.

Hint: low-temperature physics.

I have on occasion made one-turn inductors as follows:
-- Get a piece of thick-walled OFHC copper pipe, roughly
2" diameter and 4" long. Actual size may be determined
by the size of your magnet (see below).
-- Attach a plastic or wooden handle, 2 or 3 feet long,
with the axis of the handle perpendicular to the axis
of the pipe. The whole think looks sorta like a hammer.
-- Soak the pipe in liquid nitrogen. This is plenty cold
enough to eliminate the resistivity due to thermal phonon
scattering; all that's left is impurity scattering. It's
easy to find OFHC copper with an RRR (residual resistivity
ratio) in excess of 100. So you wind up with L/R times
of a second or more.
-- If you can get access to a biiig magnet (high field and
large gap) this makes an impressive demonstration. Put
the pipe into the field. You can wiggle it around freely
so long as you move it parallel to the field lines. And
you can turn it slowly. But if you try to turn it quickly,
so that the pipe turns end-for-end, reversing the flux
linkage, it will take all your strength and then some.

Also:

*) If you really want to go wild, you can use a pipe made of
superconducting material, but that's a lot more bother.

*) If you want to make an electrical circuit, get some insulated
OFHC copper wire, wind your own inductor, and soak it in
liquid nitrogen before use. (Ordinary off-the-shelf wire
is not OFHC and has an underwhelming RRR.)

======

So I'm willing to settle for having to using an oscilloscope to catch
the action. But now my question is what kind of switch to use?
Conservatively estimating an inductance of 0.1 H and a resistance of
one kilo-ohm (to allow use of an external variable resistor that
isn't dominated by the inductor's resistance), it seems to me that I
need to be able to switch cleanly on the timescale of about 10
microseconds. My experience suggests this is better than what
ordinary mechanical switches can do, which are subject to electrical
"bounces" upon contact. (I'm sure there's a technical name for this,
but you probably know what I mean.)

The technical term for "switch bounce" is "switch bounce" :-)

a) If you shop around you may be able to find a mercury-wetted
reed switch e.g.
http://www.ceejay.com/ales02.html
which is bounceless. This contains vastly less mercury than
an out-and-out mercury switch, just enough to wet the contacts.
Surface tension etc. in the liquid maintains the connection
even if the nonliquid parts bounce a bit.

b) You can get high-power FETs for cheap. Apply a goodly gate
voltage and you'll have an impressively low "on" resistance.

Be sure to use a diode and/or a properly-sized resistor in
parallel to protect it when it goes to the "off" state.