Re: Why would stranded wire cause more signal attenuation than solid?

Sleepy Hollow?  On Irondequoit Bay?  No matter.  Your question, Mr. Lohr, has 
an unwarranted assumption.  It is not that the wire is stranded that increases 
attenuation -- it's the frequency range being propagated and the mechanism for 
attenuation at that frequency.

For direct current (zero frequency), electron flow is uniform thoughout a 
conductor.  If the signal is AC, as the frequency range goes up, the electron 
flow begins to move (by magnetic effects) toward the conductor surface.  At 
high frequencies, almost all the flow is on the surface.  This is called "skin 
effect".

When skin effect is strong, most of the conductor cross-section is wasted as 
far as electrical conduction is concerned.  This increases the wire's 
"effective resistance."  It occurred to someone that if current flows in the 
skin, let's provide lots of skin.  So stranded wire with EACH STRAND INSULATED 
FROM THE OTHERS BY VARNISH was invented to see if it had a lower effective 
resistance.  It did -- over a limited frequency range like a few Megahertz.  
This wire is called Litz wire. 

But the wire you are talking about is stranded with the strands uninsulated.  
The stranding is provided for mechanical flexibility -- nothing electrical is 
involved.  The loss mechanism at 350 MHz is due to the insulating material 
surrounding the wire; that is, the dielectric.  Ideally, dielectrics have zero 
conductance (G) but practical ones don't at high frequencies.  The rapidly 
reversing voltage induces heating in the dielectric, and this creates loss.  
Recognize at 350 MHz, we are just a tad lower than microwave frequencies.

You asked about characteristic impedance and I'll tell, courtesy of Javid and 
Brenner, Analysis, Transmission, and Filtering of Signals, McGraw-Hill, 1963.  
In LaPlace domain, characteristic impedance, Z0 = SQRT((R + sL)/(G + sC)) where 
parameters R, L, G, and C are DISTRIBUTED parameters.  For example, L would per 
in Henrys per meter.  For sinusoidal steady state, s = j*2*PI*f (there's your 
tie-in to frequency).

Now here's the punchline.  For distortionless transmission, (R/L) must equal 
(G/C).  It's the wire manufacturer's job to make that so.  You must accept some 
attenuation for reasons implied before.  What the Cat 5 wire people promise is 
voltage pulses to propagate along the wire without being spread out (that's 
called "dispersion").

Hope that helps in Sleepy Hollow.  Watch out for the Headless Horseman.

Larry Skarin.