William Chops Westfield writes:
> It takes MUCH more current than 500mA to melt a piece of 30g wire-wrap wire.
> I just used my bencchtop supply to run 3A through a piece and it didn't even
> glow or discolor due to heating.
>
> That doesn't mean the rule of thumb is invalid, though.  I'd expect the
> first concern to be "excessive heating", just due to system issues. (and
> there's a LOT of wire in a transformer, so even a little bit of heat/cm will
> amount up.  Secondly, you'd run into "don't melt the insulation", followed
> by "don't melt the solder", with actual wire-melting WAY up there on the
> list of concerns.

According the the book "Reference Data for Engineers; Radio, Electronics,
Computer, and Communications, 7th ed." ISBN 0-672-21563-2 (A copy should
be on every electrical engineer's desk), the fusing current for 30 ga
copper wire is 10.2A (that is, the current at which the wire will melt).
the formula is as given I=K*d^(3/2), where d is the diameter in inches and
K is a constant that depends on the metal, aluminum K=7585, copper K=10244,
and so on.

The most important thing to be converned about is the wire (and junction)
resistance. a piece of another chart for annealed copper follows:
gauge       ohms/1000ft at 20C
12          1.588
14          2.525
18          6.385
26          40.81
30          103.2

So for power supplies, you can use ohms's law to figure out the voltage drop,
but for CMOS, it is very easy to get in trouble, because CMOS can cause
extremely high current spikes, which is why decoupling capacitors are
obligatory.  So the short answer is: for currents that pics can draw, 30ga
wire is fine.  If you need a big slug of current delivered in a short time,
start looking to heavier wires.

Why do I qualify the above paragraph for power supplies?  Remember, a wire
is not just a wire.  It is also an inductor and part of a capacitor, so
you must be aware of that on any line you are doing switching (i.e. introducing
high frequency components.

Matt








-----------------------------
Matt Bennett                |
mjb@hazmat.com              |
http://www.hazmat.com/~mjb/ |