Mike Speed wrote:

Roy Lewallen had written:
When a bundle of wires ducks under another in the direction of
current flow, the current has to migrate to the outside again, snip

There's no question that it happens


Books, and to a lesser extent the web. Information about this is
snip
I have a great deal of respect for his experience, measurements, and opinions

Again, interesting, but what's been outlined so far is not scientific.
For something of this nature to be of any utility, it must be grounded
in science.


The skin effect is most thoroughly grounded in science. What you seem to
be unaware of is that it's *so* well-known that, in any discussion about
RF engineering, the scientific proof of its existence can be 'taken as
read'.

For a detailed scientific proof of the skin effect, try:
http://tinyurl.com/brpq6

That proof is more general - and hence more powerful - than the ones you
find in most engineering texts such as Terman. It demonstrates that, if
an RF current is flowing across *any* conducting surface (not restricted
to any particular shape or cross-section) and also for *any reason* (not
limited to any particular kind of circuit or device) then there will be
a skin effect.

That's the science of it; now back to the engineering.

What Roy said was quite correct. Braid is a kind of composite conducting
surface, made up of the exposed surfaces of the individual strands. The
skin effect means that the outside of the composite surface must always
carry the highest RF current density (amperes per square micron of
cross-sectional area). So whenever the weave of the braid makes an
exposed strand dive below the surface, the RF current must cross over to
the next touching strand that is still exposed. A little way further
along the braid, it will have to cross over again... and again, and
again.

It is hard to visualize exactly how these crossovers happen on a
microscopic scale, but the physics of the skin effect dictate that it
*must* happen somehow. Obviously physical and electrical contact between
the two strands is required. We also know that electrical contact works
better when there is a strong force pushing the two conductors together,
because the force deforms the two surfaces into each other, to give a
greater contact area.

The key fact is that the contact forces between strands in a braid are
very small and unreliable. That means the RF resistance of a length of
braid will be significantly higher than for a smooth conductor with the
same external surface area.

Then it gets worse. Even the thinnest film of corrosion can disrupt the
contact between copper strands in a braid. Unless the current density is
large enough to break down this film, it means the RF current is forced
to flow into the interior of the braid. Again the exact geometry is hard
to visualize, but again the physics dictate that if an isolated
'filament' of current is forced to flow beneath a conducting surface,
the voltage drop per unit length must increase - in other words, the RF
resistance must increase.

Scientific deduction has told us that all these effects must exist.
Whatit cannot tell us is how big they are in real braid, or how
important they are in practice. For that we'll need some measured
numbers.

You have two choices he either look for existing measurements from
people who have demonstrated their competence and scientific approach;
or do it yourself.



--
73 from Ian GM3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek