On Wed, 6 Jul 2005 21:23:52 -0400, "Walter Maxwell"
wrote:


If you like, I am saying your approach is valid for lossless lines, it
is also valid for all distortionless lines, but I think it is not
accurate for lines in the general case because it isn't correct if
Xo!=0.

Owen

Owen, if X = 0 there is no attenuation, but you're saying my material is
invalid if X is not 0? I'm sorry, but I'm confused.

Walt, it has just occurred to me that I am using the "actual" Zo, not
the nominal Zo, and I think your rho calc is based on the nominal Zo,
as it will be measured with an instrument presumably calibrated for
nominal Zo.

I have compared the loss calculated by your method (with rho based on
nominal Zo, Zo=Ro+j0) and my method and they are very similar (though
not the same). I have added a function to calculate the loss using
your formula based on nominal Zo and plotted it, along with the
difference to the power based loss calc. They are at
http://www.vk1od.net/temp/reflection.htm .

If your method is based on nominal Zo, rather than the actual Zo, it
is likely to be an approximation, though on this example, it is pretty
close and probably is quite adequate for most practical lines at HF
and above. (The error increases as frequency is reduced (Zo departs
more from nominal Zo).)

Having resolved the apparent inconsistency... I am still in search of
a derivation of the Michaels formula.

Owen
--