On Wed, 9 Jun 2004 13:40:29 -0400, "Jack Painter"
wrote:

"Gary Schafer" wrote

Jack, I don't know what you have been reading in regards to skin
effect but it is very real and present.

Hi Gary, when a poster asked for the formulas for this discussion, I could
not display them in the newsgroup (ascii) so I pasted several of them on a
website.....

http://members.cox.net/pc-usa/station/skineffect.htm


I don't know what you mean "there is no standard depth for any
frequency"? It is well known.

The resistance of a particular conductor, not just it's material, must be
known to calculate skin depth. Averaging it with constants will produce the
wide variety of depths that are seen in different formulas and tables.

Yes it depends on the shape too. A round conductor will be slightly
different than a flat conductor but for our purposes it is in the ball
park. The constant comes from actual calculations. The constant makes
it easier than going through all the math to obtain the constant.


At 60 hz the skin depth is around 1/3 of an inch. Very significant in
a power transmission cable. Or a lightning ground cable..
Look up any large power cable ratings and you will usually find a DC
resistance specified and an AC resistance also specified. The AC
resistance is due to skin effect.

Yes I agreed with you it is relevant only at very high power or long lengths
when inductive reactance becomes as important as DC resistance.

The AC resistance that I am referring to has nothing to do with any
reactance due to cable length. Reactance is of course another factor
that enters into the picture but AC resistance in this case is
referring to that resistance caused by skin effect. Not reactance.


Here are some figures on skin depth for copper: Skin depth (in mils) =
2.602/(sq. root of frequency in Mhz). At 1.8 Mhz it's 1.94 mils or
.00194 inches, just under 2 thousandths. It decreases as the inverse
square root of frequency so at twice the frequency it will be .707
times as deep, and half as deep at 4 times the frequency. At 29.7 Mhz
it's about half a thousandth. At 4 or 5 skin depths any additional
thickness ceases to have additional value.

Gary, the problem with using those constants is, again, it will allow you to
reduce the skin depth to nearly nothing, when in fact below a certain cross
section at HF frequencies, formula predictions for skin depth cease to be
relevant. The current, assumed to be constant, cannot continue to use less
and less cross section until it has nothing to work with. The formulas are
an approximation that allows designers to consider the resistance casued by
skin effect and use an appropriately sized conductor. For instance, I could
not use 1,000w on thin RG-8X if your application from a table using
constants was accurate. At 5 mhz there is considerable cross section of that
small diameter center conductor carrying current. That is why the center
conductors are not paper-thin hollow tubes the way the outer shield _can_
be. Do you agree?

RG-8X will get a little warm with 1000 watts on it.
The main reason the center conductors are not paper thin hollow tubes
is because of physical restraints. If your argument would hold up then
none of the hard line coax would have hollow tubing for their center
conductors. Some of it is used in extremely high power at HF as well
as UHF. Only the outer surface of the center conductor is of much
importance in conduction.

While it is true that it gets more complicated to predict actual skin
effect on a thin conductor because as said before, the current does
not completely stop at a certain depth. It decreases exponentially.
But usually 4 or 5 skin depths are sufficient for all practical
purposes.

At that depth of 4 skin depths less than 2% of the current on the
surface will be present. We use .37 as a skin depth but .368 is closer
to what it works out to. .368 x .368 x .368 x .368 = .183 or 1.83%

But I think the original argument was whether or not the same current
or any current would flow on the inside of a copper tube at HF.
It goes away quickly and can't propagate inside as explained earlier.

Regards
Gary


Best,

Jack