Re: Skin deep oxide

[Terry Fritz <twftesla-at-uswest-dot-net>]

Hi Finn,

At 11:32 PM 7/6/00 +0200, you wrote:
>All!
>
>It has been mentioned some times that alluminium is a poor conductor at
>high frequencies. Also that the oxide layer of same is thick. I disagree
>with the last statement, the oxide layer of alluminium is very thin, and
>should not be confused with the glass layer known as anodizing.

In my experience, aluminum forms a "thin" oxide layer in hours that one has
to clean away before MIG welding and such.  Old aluminum things need
considerable sanding to get through the harder oxide layer and into the
soft pure aluminum.  I would guess about 10 to 20 mils.  Far thicker than
the skin depth.

>
>But what if this oxide layer is a poor conductor of rf, then so is the
>enamel layer of copper magnet wire, and this layer certainly never kept
>us from using the wire to wind secondary`s.

If the top layer has very high resistance such as pure aluminum oxide or
enamel, then there is no current flow and no loss.  What you have to watch
out for is the transition layer were the resistance is 10 to 1000 ohms.  If
there is significant current flow in this region, the losses will be very
high.  Due to the fields and skin depth effects, the current just seems to
naturally flow in this bad high resistance transition region in aluminum.

>
>What I`m trying to say, I think, is: how _bad_ a conductor must a given
>coating, be it varnish or oxide, be, before we can assume that the
>current stays out of it, so that it can be neglected, like we surely do
>in case of it being varnish, on the other hand,  when it is oxide, we
>worry.

The top terminal has a very large area with relatively small current.  Even
if it were 100 ohms, the losses would not be significant.  However, in the
primary circuit were the current can reach thousands of amps, the I^2R
heating and losses can be dramatic.

>
>Won`t the current just avoid the oxide layer as it would enamel, and
>sink a bit deeper _in_.

The enamel is a pure insulator so the electrons cannot flow in it.
However, the poor conducting oxide layer has high enough resistance to give
serious loss but is still conductive enough to attract the current flow
into this layer.  The current density and resulting I^2R loss is the real
key.  Only the primary circuits would be of concern to us.  "I" would not
use aluminum in the primary.  Of course, virtually all of our capacitors
have aluminum plates in them and the electric utilities use aluminum for
just about everything :-)) so there are a number of if-and-or-buts to all this.

A primary coil is probably not a good place to use aluminum.  However,
small blocks and such, as may be found in a spark gap, probably do not hurt
too much to worry with.  I would avoid ferrous metals in the primary
circuit like the plague.  Ferrous metals (high permeability) have very thin
skin depths and even a small bolt could have noticeable loss.  In my new
coil, I removed all aluminum and especially ferrous materials from the
primary current path and it really did seem to make a difference.  I like
brass and copper the best.

Silver plated copper is almost the standard in big high current RF devices.
 Silver does easily tarnish but the oxide of silver (it may actually be a
sulfide?) is still electrically conductive so even tarnished silver has
very good RF conductivity.  Gold plated copper is the best and is used in
RF power transistors.  Excellent thermal, and electrical conductivity and
no oxidation.  Looks great and may not be that expensive depending on your
local plating situation.

At work we run hundreds of RF amps though silver plated copper conductors
that still need to be internally water cooled.  Aluminum would simply burn
up and any ferrous metals, even in the area, just melts and flows away.
But if the currents are very small, such effects are of little concern.
Electrostatic machines and other very low current devices can use aluminum
for everything without any concern.

BTW - At 150kHz, the skin depth of aluminum is 0.0002 inches.  The rest of
the thickness does not really contribute electrically but does carry off
heat from the conducting layer which is actually pretty important too...

Cheers,

	Terry

>
>Cheers, Finn Hammer
>
>aka:  hammer-at-wimshurst-dot-com
>