On 6/21/2011 11:13 AM, Jim Lux wrote:

It does matter. What you really don't want is something that changes.
Either always connected or always insulated is what you want.

Dead right.

This seems to be one reason why yagis at HF tend to be grounded elements
and at VHF and up tend to be insulated. At HF it's fairly easy to get a
low resistance with fairly low impedance change over a few years.
That's harder to do at VHF especially with through the boom uninsulated.
Hence through-the-boom or over-the-boom insulated at 2m and up.

I've built many 6m beams with no correction added and using simple
homebrew U clamps. I have seen little discernible difference in their
performance from predicted. Of course I shouldn't except perhaps in the
side nulls. And they seemed about right. One small confirmation was
the SWR curve also matched predicted.

Anyway, the problem with connected is corrosion. Can only slow it down.

tom
K0TAR

On 6/22/2011 7:31 PM, tom wrote:

I've built many 6m beams with no correction added and using simple
homebrew U clamps. I have seen little discernible difference in their
performance from predicted. Of course I shouldn't except perhaps in the
side nulls. And they seemed about right. One small confirmation was the
SWR curve also matched predicted.



Based on a fair number of simulations over the years, I'd say that SWR
curves are not very sensitive to pattern degradation (except in an
egregious case)..

Consider a 5 element medium gain Yagi where the currents are all roughly
equal (which is what you'd have with moderate superdirectivity) and it
has a F/B ratio of, say, 20dB.

Screw up the current in one of the elements by, say, 25% (so the overall
excitation is now off by 5%).. That's enough to turn your -20dB null
into a -10dB null (in round numbers..)

But, would you even see that in a VSWR.. say it was 1.2:1 before
(reflection coefficient = 0.09).. and now the reflection coefficient
changes by 5%.. so it's 0.0945.. That's a VSWR of 1.208.. I think you'd
have a hard time measuring that..

On 6/23/2011 11:34 AM, Jim Lux wrote:
On 6/22/2011 7:31 PM, tom wrote:

I've built many 6m beams with no correction added and using simple
homebrew U clamps. I have seen little discernible difference in their
performance from predicted. Of course I shouldn't except perhaps in the
side nulls. And they seemed about right. One small confirmation was the
SWR curve also matched predicted.



Based on a fair number of simulations over the years, I'd say that SWR
curves are not very sensitive to pattern degradation (except in an
egregious case)..

Consider a 5 element medium gain Yagi where the currents are all roughly
equal (which is what you'd have with moderate superdirectivity) and it
has a F/B ratio of, say, 20dB.

Screw up the current in one of the elements by, say, 25% (so the overall
excitation is now off by 5%).. That's enough to turn your -20dB null
into a -10dB null (in round numbers..)

But, would you even see that in a VSWR.. say it was 1.2:1 before
(reflection coefficient = 0.09).. and now the reflection coefficient
changes by 5%.. so it's 0.0945.. That's a VSWR of 1.208.. I think you'd
have a hard time measuring that..


You are correct. I have also measured gain and F/B and side patterns.
And they match predicted quite well.

One design was a 7.5 wavelength 432 EME antenna which had nice ground
noise performance in real world use. Which is very pattern sensitive.

tom
K0TAR

On 6/23/2011 8:54 PM, tom wrote:
On 6/23/2011 11:34 AM, Jim Lux wrote:
On 6/22/2011 7:31 PM, tom wrote:

I've built many 6m beams with no correction added and using simple
homebrew U clamps. I have seen little discernible difference in their
performance from predicted. Of course I shouldn't except perhaps in the
side nulls. And they seemed about right. One small confirmation was the
SWR curve also matched predicted.



Based on a fair number of simulations over the years, I'd say that SWR
curves are not very sensitive to pattern degradation (except in an
egregious case)..

Consider a 5 element medium gain Yagi where the currents are all roughly
equal (which is what you'd have with moderate superdirectivity) and it
has a F/B ratio of, say, 20dB.

Screw up the current in one of the elements by, say, 25% (so the overall
excitation is now off by 5%).. That's enough to turn your -20dB null
into a -10dB null (in round numbers..)

But, would you even see that in a VSWR.. say it was 1.2:1 before
(reflection coefficient = 0.09).. and now the reflection coefficient
changes by 5%.. so it's 0.0945.. That's a VSWR of 1.208.. I think you'd
have a hard time measuring that..


You are correct. I have also measured gain and F/B and side patterns.
And they match predicted quite well.

One design was a 7.5 wavelength 432 EME antenna which had nice ground
noise performance in real world use. Which is very pattern sensitive.

tom
K0TAR


In fact, I would assert that today, VSWR measurements are almost useless
for *antenna adjustment and construction*.. and even for adjusting
elements. With modern modeling codes which don't require all sorts of
little fiddly "correction factors" (e.g. taper or boom factors) to
allow simulating a complex structure with a simpler model, a tape
measure would be a better tool.

There's a lot of literature from the 50s-70s and even 80s where antennas
are constructed and you adjust the element length by using the trusty
GDO. (Yep, I remember helping my grandfather assemble a multiband Yagi
of some sort in the early 70s doing it that way..)

But today.. mechanical dimensions and a high fidelity model are the way
to go.


Where the VSWR measurement would be useful is in two places:
1) Adjusting a wire antenna length... they're usually interacting with
the surroundings, so you have to cut and try, and a VSWR sweep is a good
way to get there quickly. Putting up that triband multiwire dipole for
Field day tomorrow? The Antenna analyzer is your friend when trying to
get all those mutually interacting lengths dialed in.


2) A quick check to see if an antenna system that worked before has
"broken" in a big way.