[Top] [All Lists]

Re: [Antennaware] EZNEC modelling

To: David Gould <>
Subject: Re: [Antennaware] EZNEC modelling
From: Terry Conboy <>
Date: Mon, 20 Nov 2006 16:40:10 -0800
List-post: <>
At 01:34 PM 2006-11-20, David Gould wrote:
>I see what you have done,  in that the capacitor has gone to ground 
>effectively in parallel with the coil.
>My thinking is that the series C does indeed go to earth but 
>**through**  the 50ohm resistive source impedance.
>Is it just a convention that this resistance can be ignored in the 
>model?  The proof is that it seems to work without it..

The model sources in NEC are either current or voltage sources, so 
are effectively either infinite or zero impedance.  You may note that 
the gain of a modeled antenna is independent of the type of 
source.  I.e., the gain doesn't include any mismatch loss.  The 
source looks into the load of the antenna, in this case, through the 
series capacitor.

I chose to put the cap in the shorter path to ground, since a little 
more current flows through it.  It would be quite possible to put the 
inductor in the direct path to ground, the capacitor in the 
horizontal conductor and the source in the off-axis conductor to 
ground.  The results should be similar, except for the inductance due 
to the differing lead lengths.

In the real world, ham transmitters seldom present a 50 ohm output 
impedance, they are just designed to work into a nominal 50 ohm 
load.  (Much like audio amps are usually designed to work into 4 or 8 
ohm loads, but normally have a near-zero output impedance.)

>I have got TLW on a CD somewhere, I must look it out and load it 
>onto this (new) PC.

I find this is a handy program, since it designs matching networks 
with practical (lossy) components and computes the network loss 
(which can be very high for certain network configurations, such as 
high-pass T networks with small caps and a low Z, capacitive load).

>I should just say that I made a small mistake in the previous post 
>in that the vertical was only 37ft high (not 48) so that was why my 
>L-match components were different from yours.  (I have been playing 
>with several different models and got them mixed up)  I have 
>modified your model to the correct values.

That's a lot less "touchy" at around 3/8 wavelength.  I'll update my 
model, too.

73, Terry N6RY

Antennaware mailing list

<Prev in Thread] Current Thread [Next in Thread>