AVS Forum Special Member

Join Date: Nov 2006

Location: Bushnell's Basin, NY

Posts: 1,519

Mentioned: 6 Post(s)

Tagged: 0 Thread(s)

Quoted: 262 Post(s)

Liked: 150

When doing an NEC modeling exercise, the size and location of every bit of metal has to be included. However, once that's done, there is a much simpler option that retains a high degree of accuracy over a limited range - scaling.

In my case, I started with a CM4221, which when properly measured has certain relationships between physical size and frequency response. Peak gain is at a frequency whose wavelength equals the width of the bowtie when wiskers are spread. This is an observation only, but it fits with how EM radiation interacts with matter.

If one changes all dimensions when making a copy, you get a scale model of the original, and in antennas, that means you can scale the frequency axis. If it's twice the size, just divide each frequency in the gain table by 2 to get the new gain curve. (But 2x is a BIG change! Scaling may not apply this far out.)

Conversely, if you know an antenna's peak gain frequency and you want to move it, your scale factor becomes the frequency ratio. In this case, a CM-4221 peak gain is Ch 66, 785 MHz, and you want to move it to Ch 51, 695 MHz. The ratio is 785/695=1.129x, which is real close to 1 1/8 (1.125).

The key is that ALL DIMENSIONS have to change by this amount. Some things, like wire diameter, have smaller effects, but basic building blocks like element size and location have to all change by the same factor to maintain any degree of predictability.

Frank