[TowerTalk] Modeling vs Experimenting Crowds

Grant Saviers grants2 at pacbell.net
Thu Sep 12 11:17:55 EDT 2019



On 9/11/2019 9:45 PM, David Gilbert wrote:
> 
> I think most of that is fairly misguided advice when it comes to antennas.
> 
>
I would agree and the rest well said.

The NEC programs continually amaze me how well they predict behavior and 
performance.  There is very little magic left beyond Maxwell.  Creative 
ideas?  Great - give them a spin in the NEC washing machine. I must have 
200 bad ones thankfully never built.  Plus, EZNEC has some nifty 
generation routines.  Want a tapered wound spiral vertical? - easy.  A 
few hours of modeling and then we built one - resonance F and R very 
close, measured A/B gain in the predicted minus few db ballpark. Was 
that model perfect? - of course not, some significant variables were not 
possible to model, but the prediction was very close to reality.  As 
W8JI pointed out, without controlled A/B performance testing, there are 
just too many variables to declare any sort of victory.  That pretty 
much wipes out experimental optimization unless you have a real antenna 
range.

I find it depressing to see hams put up at great effort even simple wire 
antennas that just don't work very well. No need to debate which yagi is 
better.  Now when asked to help build a DOG of an antenna, I pass. 
Especially if I was asked and counseled against it.  The folklore that 
causes folks to build (and buy) them has amazing staying power.

Junk antennas are likely one reason folks get frustrated and quit the 
hobby.  So, I think we have a self interest to encourage building (or 
buying) good ones.

Grant KZ1W


More information about the TowerTalk mailing list