Hi,
I have done quite a bit of numerical analysis of VHF Yagi's with usually
very good agreement between theory and performance at least for gain and
pattern. One of the things that seems more difficult to correctly predict
however is the input impedance, and thus to design some sort of a match
without some sort of subsequent trial and error. In an effort to explore the
differences further I have been looking at ways of measuring the impedance
directly rather than indirectly by vswr with the matching in place. I have
made a number of different return loss bridges, and even tried the technique
described in an old Ham Radio Article where you take two VSWR readings with
and without an added series resistance. All of course making allowances for
coax length etc. The problem I have is all of them give, in some cases
wildly, different answers even when used on the same antenna. So my
questions a

What sort of accuracy can I expect from these sorts of methods?

Is there a better way (which doesn't involve large sums of money) to measure
antenna impedance at say 146Mhz?

Thanks

Paul VK3DIP

"Paul (Home) News" wrote
I have done quite a bit of numerical analysis of VHF Yagi's with usually
very good agreement between theory and performance at least for gain and
pattern. One of the things that seems more difficult to correctly predict
however is the input impedance, and thus to design some sort of a match
without some sort of subsequent trial and error. In an effort to explore
the
differences further I have been looking at ways of measuring the impedance
directly rather than indirectly by vswr with the matching in place. I have
made a number of different return loss bridges, and even tried the
technique
described in an old Ham Radio Article where you take two VSWR readings
with
and without an added series resistance. All of course making allowances
for
coax length etc. The problem I have is all of them give, in some cases
wildly, different answers even when used on the same antenna. So my
questions a

What sort of accuracy can I expect from these sorts of methods?

Is there a better way (which doesn't involve large sums of money) to
measure
antenna impedance at say 146Mhz?

Thanks

Paul VK3DIP

===============================
Paul,

Attempts to accurately determine antenna input impedance, using an
inherently ambiguous, innacurate SWR meter at the transmitter end of a line
of uncertain length and velocity, are doomed to failure.

*Never* expect to obtain numbers worthy of serious engineering application.
There are far too many uncertainties of unknown magnitudes.

The only way of obtaining an accurate measurement is to climb a ladder
taking with you an R plus or minus jX hand-held impedance bridge. Can you
borrow one ?

But why do you wish to know antenna input impedance when you are aleady
quite happy with using an inaccurate SWR meter to fiddle a 1-to-1 SWR at the
transmitter end.

The ultimate objective, of course, is just to obtain a 50-ohm load for the
transmitter regardless of what the SWR and antenna impedance might be.
----
Reg, G4FGQ

The input impedance can be measured reasonably well at ground level.

Align the antenna so that the reflector is 'down' and the last director
is 'up'. Ground effects are minimized due to the F/B of the antenna. The
antenna is radiating straight 'up'.

Next take a 1 wavelength, allowing for velocity factor, coax line and
connect it to the antenna feedpoint.

Finally, beg, borrow, requisition, pilfer, rustle, etc., an antenna
analyzer similar to the MFJ 259B. Connect it to the other end of the 1
wavelength coax.

Select the function to read impedance. Dial in your frequency and read
the impedance.

A one minute job once the antenna, coax and meter are at hand.

+ + +

Paul (Home) News wrote:

Hi,
I have done quite a bit of numerical analysis of VHF Yagi's with usually
very good agreement between theory and performance at least for gain and
pattern. One of the things that seems more difficult to correctly predict
however is the input impedance, and thus to design some sort of a match
without some sort of subsequent trial and error. In an effort to explore the
differences further I have been looking at ways of measuring the impedance
directly rather than indirectly by vswr with the matching in place. I have
made a number of different return loss bridges, and even tried the technique
described in an old Ham Radio Article where you take two VSWR readings with
and without an added series resistance. All of course making allowances for
coax length etc. The problem I have is all of them give, in some cases
wildly, different answers even when used on the same antenna. So my
questions a

What sort of accuracy can I expect from these sorts of methods?

Is there a better way (which doesn't involve large sums of money) to measure
antenna impedance at say 146Mhz?

Thanks

Paul VK3DIP

Hi Paul

Gordon VK2ZAB (and others) published some time ago a complex Z bridge
thing for VHF/UHF. It uses transmission lines for tuned elements and is
band specific.

Try http://www.vhfdx.oz-hams.org/measurements.html

And do other google searches using Gordons callsign

Cheers Bob VK2YQA

Paul (Home) News wrote:
Hi,
I have done quite a bit of numerical analysis of VHF Yagi's with usually
very good agreement between theory and performance at least for gain and
pattern. One of the things that seems more difficult to correctly predict
however is the input impedance, and thus to design some sort of a match
without some sort of subsequent trial and error. In an effort to explore the
differences further I have been looking at ways of measuring the impedance
directly rather than indirectly by vswr with the matching in place. I have
made a number of different return loss bridges, and even tried the technique
described in an old Ham Radio Article where you take two VSWR readings with
and without an added series resistance. All of course making allowances for
coax length etc. The problem I have is all of them give, in some cases
wildly, different answers even when used on the same antenna. So my
questions a

What sort of accuracy can I expect from these sorts of methods?

Is there a better way (which doesn't involve large sums of money) to measure
antenna impedance at say 146Mhz?

Thanks

Paul VK3DIP

Dave, I was replying to the original questioner. But by immediately
following your response with mine and including a comment of yours caused a
little confusion. Sorry!

I agree your method will work. The problem, a practical one, is obtaining a
COAXIAL line length exactly an integral number of 1/2-wavelengths long.
There's no way of proving it exept by climbing a ladder and disconnecting
the line from the antenna.

And it is an exact 1/2-wavelength long at ONE frequency only. But it is
required to make measurements over a whole band of frequencies. To shift to
other frequencies
involves calculations taking Zo into account. But Zo is not accurately
known. So then you have to measure line Zo. And so on.

And you have to know exactly what you are doing because the 259B does not
provide the sign of jX in R+jX.

But as I said before, all you want to know is whether or not the transmitter
is loaded with 50 ohms. To hell with SWR and antenna input impedance. ;o)
----
Reg, G4FGQ

Thanks Bob (Bob)

I'm going to make one of those impedance "meters". I sure appreciate
having guys like you do all the research work for me. Thanks again.

Jerry




"Bob Bob" wrote in message
...
Hi Paul

Gordon VK2ZAB (and others) published some time ago a complex Z bridge
thing for VHF/UHF. It uses transmission lines for tuned elements and is
band specific.

Try http://www.vhfdx.oz-hams.org/measurements.html

And do other google searches using Gordons callsign

Cheers Bob VK2YQA

Paul (Home) News wrote:
Hi,
I have done quite a bit of numerical analysis of VHF Yagi's with usually
very good agreement between theory and performance at least for gain and
pattern. One of the things that seems more difficult to correctly
predict
however is the input impedance, and thus to design some sort of a match
without some sort of subsequent trial and error. In an effort to explore
the
differences further I have been looking at ways of measuring the
impedance
directly rather than indirectly by vswr with the matching in place. I
have
made a number of different return loss bridges, and even tried the
technique
described in an old Ham Radio Article where you take two VSWR readings
with
and without an added series resistance. All of course making allowances
for
coax length etc. The problem I have is all of them give, in some cases
wildly, different answers even when used on the same antenna. So my
questions a

What sort of accuracy can I expect from these sorts of methods?

Is there a better way (which doesn't involve large sums of money) to
measure
antenna impedance at say 146Mhz?

Thanks

Paul VK3DIP

I want to caution you about using a half or one wavelength line to do
measurements. That's a viable method if the imedance of the antenna is
close to the characteristic of the line. If it's not, you'll find that
even a surprisingly small line loss -- one that you'd normally consider
negligible, can seriously skew your results. The calculation is
straightforward presuming you know the loss -- I'm sure Reg's program
would be adequate. Do some what-ifs with various antenna impedances and
you'll see what I mean. The effect gets worse as the antenna and
transmission line impedances get more different, and as the line gets
longer. That is, a one wavelength line will have more effect than a half
wavelength one.

Also, line length becomes more and more critical as the impedance of the
antenna and transmission line become more different and as the line gets
longer. Again, a little experimentation with the calculations will
illustrate what to expect.

Even if you carefully account for the transformation of the connecting
line or don't use any line at all, you have to be aware of common mode
currents and how your test setup differs from your normal rig
connection. And finally, even with a perfect lab setup, you'll find that
good impedance measurements can be hard to make with amateur equipment.

Before you get carried away, make some measurements on the bench with
your meter and using good quality loads, or at least RC combinations
using chip resistors and capacitors or ones with extremely short leads.
Make impedances similar to ones you hope to measure. If you can get
values which are accurate enough to suit you, go to the next step and
measure the same loads through a transmission line as has been
suggested, and see if you're able to extract the actual load value from
the measured value with sufficient accuracy.

If you get that far, you've partially answered your question about what
kind of accuracy to expect, and you're ready to start figuring out how
to deal with common mode currents.

Decent antenna impedance measurements aren't simple to make, even at HF.
They're more difficult at VHF and above.

Roy Lewallen, W7EL

Reg Edwards wrote:

Dave, I was replying to the original questioner. But by immediately
following your response with mine and including a comment of yours caused a
little confusion. Sorry!

I agree your method will work. The problem, a practical one, is obtaining a
COAXIAL line length exactly an integral number of 1/2-wavelengths long.

The MFJ 259B will measure the length of the line for you before test.
Leave it open circuited and connect the other end to the MFJ. It finds
the 1/2 wavelength frequency.

There's no way of proving it exept by climbing a ladder and disconnecting
the line from the antenna.

A bench setup with a 1 wavelength line does not require climbing and
measuring at the tower.


And it is an exact 1/2-wavelength long at ONE frequency only. But it is
required to make measurements over a whole band of frequencies. To shift to
other frequencies
involves calculations taking Zo into account. But Zo is not accurately
known. So then you have to measure line Zo. And so on.

Agree it is a one frequency measurement.


And you have to know exactly what you are doing because the 259B does not
provide the sign of jX in R+jX.

But, the sign of X is very easy to determine. Increase frequency
'slightly' on the 259B and observe absolute value of X. If X increases
it is inductive. If X decreases it is capacitive. BTW the Or is a value
at antenna resonance only at one frequency. But, you a very well aware
of that. The comment is made for other readers.


But as I said before, all you want to know is whether or not the transmitter
is loaded with 50 ohms. To hell with SWR and antenna input impedance. ;o)
----
Reg, G4FGQ

There is one possibility remaining. If the Yagi is to be tuned for
MAXIMUM gain, and that is the objective, then Ro will be the lowest
value at resonance. Most Ham yagis are not tuned for optimum gain as we
all know.

There is WAY WAY too much emphasis among today's hams regarding low
VSWR. My 75/80 doublet has a VSWR approaching 30:1 and works just FB!!
The problem is the absence of output tuning in most of the rigs
available today. Oh! for my OLD VIKING II or my Drake 4C. grin

Gor Bless and you have my permission to celebrate the USA Father's Day.

Roy, your caution is well placed.

As I've said before, most amateurs and professionals (it seems from these
walls) suffer from delusions of accuracy.

Their delusions are seldom frustrated by things going wrong after some
tedious, supposedly highly accurate, design work has been done. They
congratulate themselves on their success and sometimes follow up by writing
learned papers incorporating 6 places of decimals about it.

But in engineering reallity, especially with Radio, things work simply
because any bloody thing will work after a fashion. And if the transmitter
is loaded with an actual but unknown impedance between 30 and 80 ohms, such
that it works, they continue to remain oblivious to their delusions.

I relate this, certainly not to cast ridicule, but with the inention of
enhancing the underlying beauty of this intriguing hobby of ours.

Perhaps 'suffer from' are the wrong words.

In the UK it is Father's Day. So I am about to pour myself another glass
from a bottle of special reserve port, a thoughtful present from a loving
'doter'.
----
73 and 88, Reg, G4FGQ




"Roy Lewallen" wrote in message
...
I want to caution you about using a half or one wavelength line to do
measurements. That's a viable method if the imedance of the antenna is
close to the characteristic of the line. If it's not, you'll find that
even a surprisingly small line loss -- one that you'd normally consider
negligible, can seriously skew your results. The calculation is
straightforward presuming you know the loss -- I'm sure Reg's program
would be adequate. Do some what-ifs with various antenna impedances and
you'll see what I mean. The effect gets worse as the antenna and
transmission line impedances get more different, and as the line gets
longer. That is, a one wavelength line will have more effect than a half
wavelength one.

Also, line length becomes more and more critical as the impedance of the
antenna and transmission line become more different and as the line gets
longer. Again, a little experimentation with the calculations will
illustrate what to expect.

Even if you carefully account for the transformation of the connecting
line or don't use any line at all, you have to be aware of common mode
currents and how your test setup differs from your normal rig
connection. And finally, even with a perfect lab setup, you'll find that
good impedance measurements can be hard to make with amateur equipment.

Before you get carried away, make some measurements on the bench with
your meter and using good quality loads, or at least RC combinations
using chip resistors and capacitors or ones with extremely short leads.
Make impedances similar to ones you hope to measure. If you can get
values which are accurate enough to suit you, go to the next step and
measure the same loads through a transmission line as has been
suggested, and see if you're able to extract the actual load value from
the measured value with sufficient accuracy.

If you get that far, you've partially answered your question about what
kind of accuracy to expect, and you're ready to start figuring out how
to deal with common mode currents.

Decent antenna impedance measurements aren't simple to make, even at HF.
They're more difficult at VHF and above.

Roy Lewallen, W7EL

Dave Shrader wrote:
If the Yagi is to be tuned for MAXIMUM gain, and that is the objective,
then Ro will be the lowest value at resonance.

That's an interesting assertion. Do you have further evidence for it?


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek