[TowerTalk] OWA Yagis was 40m 4el KLM - replacing linear loading with coils

[Richard (Rick) Karlquist]

On 4/30/2020 4:15 PM, sawyered at earthlink.net wrote:
> In my view, the discussion on OWA yagis is about the element count and the

> I have all conventional yagis.  The 2:1 bandwidth of the my yagis are 28000
> - 28750.  2100 - 21375.  14000 - 14300.  7000 - 7185.  The amp is at full
> power up to 3L1 SWR.  And I have most of my runs using LDF5-50.  So in my
> opinion, for a given boom length, the trade of the OWA isn't worth the gain
> loss outside the 2:1 edges of the band.  Especially because I am almost
> never running in those sections of the band - where gain is most important.

Unfortunately, the pattern of a Yagi can go all to hell as you move away
from the design center frequency, no matter what your ability of match
it.  The "2:1 VSWR bandwidth" doesn't tell you much about how the 
pattern holds up across frequency.

I can easily illustrate this with my MonstIR.  If I set it up for the
bottom end of 40, then drive it at the top end of the band, I can
use the driven element tuning (independent of the parasitic elements)
to somewhat mitigate the VSWR.  But the pattern degrades considerably
and I lose several dB or more of gain.


> Like everything, there is no free lunch in engineering trade-offs.

At least in terms of performance ONLY, the SteppIR is a free lunch.
AFAIK, it beats anything else in terms of gain across the entire
band, given a boom length and number of elements.  Yes, there are
various non-performance tradeoffs, but a properly working one is
tough to beat.  YMMV, but my MonstIR has worked perfectly for over
10 years.  BTW, the SteppIR driven element has an impedance of
22.2 ohms, and the loss in the transformer/balun is extremely low.  I 
measured in on a VNA before I put up the antenna.

There are alternate designs that emulate the SteppIR in a monobander
by individually tuning the elements using various modalities, such
as motor tuned or relay switched capacitors or inductors.  They
produce similar results if properly implemented.

> Ed  N1UR
> 

Rick N6RK