At 12:17 AM 11/14/2003, Daniel L. Katz wrote:
>the gap drops from infinity to probably a few hundred ohms at the peak of
>the spark current, while the voltage across the gap drops, respectively,
>from 10,000 V to a roughly a few to several tens of volts (with most of
>the voltage drop across the coil secondary, and coil and plug wires).

I'm guessing that it remains a good bit higher since we're dealing with a
spark (vs an arc which can carry heavy current at low voltages by
transferring metal ions across the gap).

>complicated. the resistance of the spark gap itself depends on the current
>flowing, temperature, pressure, and other factors. ohms law doesn't apply
>to gaseous conduction.

I don't see how this bothers Ohm's Law -- it talks about results, not
causes.  And since it only considers R and not L or C, it applies to
steady-state conditions where they have no effect, or circuits where L and
C are too small to have a significant effect under the given conditions --
or to an infinitesimal instant where you can measure instantaneous voltage
and current.  It works perfectly well to characterize an average effective
resistance or impedance in any steady-state AC system, taking L and C into
account as modifying factors applied to R, resulting in Z.

>what about the resistance of a copper wire carrying a spark current? even
>ignoring reactive effects, since the current is pulsed, the the effective
>cross sectional area of the wire is reduced by the skin effect in a
>complicated time varying manner.

Er...that *is* a reactive effect, no?  But at any instant the cross-section
will be effectively some specific value, and the relationship (as you say,
ignoring L and C) will hold.

>then we have our coolant level electrodes. note that that system uses AC
>as a dodge around the failure of ohms law for DC.

Actually it doesn't -- it applies a high-impedance square-wave (i.e. pulsed
DC, CMOS levels if I remember right) trigger signal to the alarm output
stages, and at the same time to one  of the sender electrodes which is
pulled to a static level (ground as it happens, but the circuit doesn't
care) by conductance through the coolant to the other electrode.  As long
as the signal voltage stays at a steady state the coolant alarm signal is
inhibited; if the square wave appears it means the connection between the
electrodes has been broken, and the alarm triggers.

It's a clever arrangement and I haven't thought through exactly why they
did it this way; but I don't see what failure of Ohm's Law has to do with
it.  Or what the failure is...

>diodes are useful because they don't obey ohms law.

Diodes and other active components have qualities beyond resistance (and
capacitance and inductance).  Ohm's Law doesn't address such things -- but
they do result in an "effective resistance" under a given set of conditions
that can be characterized by Ohm's Law.  Also, no device "obeys" Ohm's Law
-- it's the relationship between I and E and R that obey...I really think
you're barking up the wrong tree here.

cheers,
david

--
David Beierl -- dbeierl@attglobal.net