[TowerTalk] Lossy traps?

[Joe Subich, W4TV]

> Why would it be difficult?  You run the model with different 
> FR cards, you look at the loss number. 

Because the loss (trap Q) is not constant.  If you are willing 
to rebuild your models from band to band it is possible to 
get close but optimization can be difficult.  The change in 
Q effects both current level and phase which significantly 
impact both efficiency and directivity.  

> What would be interesting is to see the "rest of the pattern" 
> to understand the gain variation.  Is it a loss issue or a 
> less directivity issue? 
 
The published pattern data is relatively sparse but my feeling 
is that within the measurement resolution the gain variation 
appears to be primarily due to lower efficiency.  

73, 

   ... Joe, W4TV 
 

> -----Original Message-----
> From: jimlux [mailto:jimlux at earthlink.net] 
> Sent: Thursday, October 30, 2008 9:55 AM
> To: Joe Subich, W4TV
> Cc: 'Tower Talk'
> Subject: Re: [TowerTalk] Lossy traps?
> 
> 
> Joe Subich, W4TV wrote:
> > 
> >> I'm sure someone has done a nice NEC analysis comparing the
> >> two, at least from a loss standpoint. I'll bet just about 
> >> anything is less than 10% loss, so even if you halved it 
> >> to 5%, the forward gain is going to change a tiny fraction 
> >> of a dB. You might do better just wrapping the elements in 
> >> copper foil.
> > 
> > It's rather difficult to do an analysis for more than two
> > bands due to the various interactions and changing loss 
> > with frequency due to circulating currents, etc.  However, 
> > I believe L.B. Cebik, W4RNL had a fairly good analysis for 
> > a two band trapped yagi on his web site.  
> 
> Why would it be difficult?  You run the model with different 
> FR cards, 
> you look at the loss number. We run NEC models at work over a 
> decades of 
> bandwidth all the time, although some care is needed in building the 
> model and dealing with the segmentation. Traps are a problem because 
> modeling them is tricky.. they have both distributed and lumped 
> characteristics, perhaps that's what you're referring to?
> 
> 
>   Summarizing the results can be tricky (since what is "good" 
> is in the 
> eye of the beholder in many cases)
> 
> 
> 
> > 
> > Practical measurements using the K7LXC/N0AX protocol tend
> > to show that the two element per band Force 12 designs tend 
> > to show as much "real world" gain as larger (e.g. TH7DXX 
> > and KT36XA) trapped designs.  The larger "multi-monoband" 
> > designs (e.g., C31XR) will have somewhat higher gain). 
> > The tunable antennas (e.g. SteppIR) will generate higher 
> > gain than any multiband antenna of the same boom length 
> > with a three element SteppIR "real world" gain within 
> > 10% of longer boom "multi-monoband" designs (where the 
> > full boom is not used on every band) and the four element 
> > SteppIR showing "real world" gain 50 to 60% higher than 
> > a comparable length "multi-monoband" design. 
> 
> What would be interesting is to see the "rest of the pattern" to 
> understand the gain variation.  Is it a loss issue or a less 
> directivity 
> issue? Or a trading bandwidth for gain? I'll have to dig out 
> the report 
> and look at it again.
> 
> 
> > 
> > Analysis of data from several different tests/testers
> > all using the same protocols indicate a gain deficit 
> > (measured vs. expected) of 1-2 dB for trapped antennas.     
> >