"Dr. Slick" wrote in message
om...

If you measure the SWR at point 1, you let Z1 (Zload) be 50 ohms,
and Zo is the output impedance of the PA. SWR meter1 should be
characterized to Zo.
snip
Reflection Coefficient looking into load = (Zl-Zo)/(Zl+Zo)

It should be easy to see from this well known equation that a
perfect 50 Ohm Zl will definitely change SWR as you change Zo away
from 50 ohms.


Slick

Looking at a book. Has the same equation, only thing is that Zo is the
impedance of the *coax*.

Tam/WB2TT

"Dr. Slick" wrote in message
. com...

If you measure the SWR at point 1, you let Z1 (Zload) be 50 ohms,
and Zo is the output impedance of the PA. SWR meter1 should be
characterized to Zo.
snip
Reflection Coefficient looking into load = (Zl-Zo)/(Zl+Zo)

It should be easy to see from this well known equation that a
perfect 50 Ohm Zl will definitely change SWR as you change Zo away
from 50 ohms.


Slick

Looking at a book. Has the same equation, only thing is that Zo is the
impedance of the *coax*.

Tam/WB2TT


Dang it, Tam, you just had to tell him! I was just having a good laugh when
you
had to go and spoil it.
73,
Tom Donaly, KA6RUH

(Tdonaly) wrote in message ...

Looking at a book. Has the same equation, only thing is that Zo is the
impedance of the *coax*.

Tam/WB2TT


Dang it, Tam, you just had to tell him! I was just having a good laugh when
you
had to go and spoil it.


Zo doesn't have to be the coax impedance. It can be any impedance
you want. What about interstage-matching networks between FETs? Do
you think there are no reflections there? Should you attempt to match
impedances there? I would think so.

let me rephrase:

PA--+SWR meter1+--50ohm line--+SWR meter2+--50ohm line--+50 ohm
(same Z as PA) (50 Ohms) dummy
load
1 2 3


If you measure the SWR at point 1, you let Z1 (Zload) be 50 ohms,
and Zo is the output impedance of the PA. SWR meter1 should be
characterized to Zo.

I've agreed that SWR meter2s SWR reading (at point 3) will not
change as you change PAs impedance, but the SWR at point 1 definitely
WILL.

So, Does the SWR change as you change the source impedance? IT
DEPENDS ON WHERE YOU MEASURE IT! Right at the output of the source,
it will!

From Pozar's Microwave Engineering (Pg. 606):

Reflection Coefficient looking into load = (Zl-Zo)/(Zl+Zo)

It should be easy to see from this well known equation that a
perfect 50 Ohm Zl will definitely change SWR as you change Zo away
from 50 ohms.


Slick

let me rephrase:

PA--+SWR meter1+--50ohm line--+SWR meter2+--50ohm line--+50 ohm
(same Z as PA) (50 Ohms) dummy
load
1 2 3


If you measure the SWR at point 1, you let Z1 (Zload) be 50 ohms,
and Zo is the output impedance of the PA. SWR meter1 should be
characterized to Zo.

I've agreed that SWR meter2s SWR reading (at point 3) will not
change as you change PAs impedance, but the SWR at point 1 definitely
WILL.

So, Does the SWR change as you change the source impedance? IT
DEPENDS ON WHERE YOU MEASURE IT! Right at the output of the source,
it will!

From Pozar's Microwave Engineering (Pg. 606):

Reflection Coefficient looking into load = (Zl-Zo)/(Zl+Zo)

It should be easy to see from this well known equation that a
perfect 50 Ohm Zl will definitely change SWR as you change Zo away
from 50 ohms.


Slick

"Dr. Slick" wrote in message
om...
(Tdonaly) wrote in message
...
Zo doesn't have to be the coax impedance. It can be any impedance
you want. What about interstage-matching networks between FETs? Do
you think there are no reflections there? Should you attempt to match
impedances there? I would think so.

You transform the input impedance of the driven stage to the optimum load of
the driving stage. Whether that results in a conjugate match is completely
immaterial. Trust me, I have done it .

As for the Zo thing, you are trying to misapply the formula.

BTW, the one place where you would think conjugate matching should always
apply, antenna to receiver input, does not (always). Lowest noise figure
often occurs at SWRs around 1.5

Tam/WB2TT

Why don't you try it? It is easy enough to add impedance between the
transmitter and SWR meter. Keep the power low, but exactly the same for the
two cases. In fact, all you would have to do is replace the 50 ohm line
between the transmitter and meter with 1/4 WL of 75 Ohm line.

Tam/WB2TT
"Dr. Slick" wrote in message
m...
let me rephrase:

PA--+SWR meter1+--50ohm line--+SWR meter2+--50ohm line--+50 ohm
(same Z as PA) (50 Ohms) dummy
load
1 2 3


If you measure the SWR at point 1, you let Z1 (Zload) be 50 ohms,
and Zo is the output impedance of the PA. SWR meter1 should be
characterized to Zo.

I've agreed that SWR meter2s SWR reading (at point 3) will not
change as you change PAs impedance, but the SWR at point 1 definitely
WILL.

So, Does the SWR change as you change the source impedance? IT
DEPENDS ON WHERE YOU MEASURE IT! Right at the output of the source,
it will!

From Pozar's Microwave Engineering (Pg. 606):

Reflection Coefficient looking into load = (Zl-Zo)/(Zl+Zo)

It should be easy to see from this well known equation that a
perfect 50 Ohm Zl will definitely change SWR as you change Zo away
from 50 ohms.


Slick

"Tarmo Tammaru" wrote in message ...
Why don't you try it? It is easy enough to add impedance between the
transmitter and SWR meter. Keep the power low, but exactly the same for the
two cases. In fact, all you would have to do is replace the 50 ohm line
between the transmitter and meter with 1/4 WL of 75 Ohm line.



You mean change the coax from point 1-2? that wouldn't be the same
as changing the PAs output impedance and the impedance of the SWR
meter1.


Slick

"Ian White, G3SEK" wrote in message ...

Now let's simplify it even further:
1. PA
2. Either SWR meter 1 or SWR meter 2.
3. 50 ohm load.



You would then have either:

PA--+SWR meter1+50 ohm dummy load
(same Z as PA)

1 2


In which case, this would be equivalent to re-normalizing the
Smith, and as you change the Z of both the PA and SWR meter, you will
absolutely measure a different SWR (you will measure the reflected
power at point 2), going into 50 Ohms believe it or not. Please get
MIMP, it's a free program from Motorola.


Or:

PA--+SWR meter1+50 ohm dummy load
(50 Ohm meter)

1 2


In which case, you would NOT measure the reflected power at point 1
(from the discontinuity from the PA to 50 ohms, because this is before
the meter.
This is the case you guys are talking about, WHICH I TOTALLY AGREE
WITH, as the source Z shouldn't affect the SWR meter in this case.

The point is that it depends on where you measure it!




As Pozar - and countless others - correctly state, rho is a function of
both the Zl that you're measuring, and the system reference impedance
Zo. You cannot either measure or calculate rho without first assuming a
value for Zo; not ever.

Since you have to choose a value of Zo *before* you can calibrate the
rho-meter, it's not surprising that two differently-calibrated meters
read differently when connected to the same physical load impedance.
It's also not surprising that "re-normalizing" the Zo, either on the
Smith chart or in a computer program, will predict exactly what *both*
meters should read. You only need to know Zl and the two assumed values
of Zo for which the respective meters were calibrated.


The point is: Zo doesn't have to be 50 ohms necessarily.


The same would be true even if the terminating impedance Zl was not 50
ohms as assumed above. The measured values would alter, but the ability
to predict the readings of both meters would not.

In that respect, you've only been stating the obvious all the time.


And i've been told i was wrong from the start! But this is typical
of a NG, isn't it.



The other part, the apparent belief that a rho- or "SWR" meter will
change its readings according to the actual, physical source impedance
from which it's being fed, remains simply incorrect. The meter is only
affected by the impedance it sees at the downstream (output) side.


agreed, but if the meter is the same Z as the PA, you will indeed
measure the discontinuity going to 50 Ohms.


Slick

Dr. Slick wrote:
"Ian White, G3SEK" wrote in message
...

Now let's simplify it even further:
1. PA
2. Either SWR meter 1 or SWR meter 2.
3. 50 ohm load.



You would then have either:

PA--+SWR meter1+50 ohm dummy load
(same Z as PA)

1 2


In which case, this would be equivalent to re-normalizing the
Smith, and as you change the Z of both the PA and SWR meter,

When you talk about "the Z of the PA", you mean the output impedance.
When you talk about "the Z of the SWR meter", you mean the system
reference impedance Zo for which the meter is calibrated. These are two
completely different physical quantities. It is totally unrealistic to
talk about varying those two quantities together, keeping them
numerically equal like they were on some kind of twin-gang pot.

However, it's not completely beyond the reach of imagination, so let's
imagine for the sake of argument that somehow you could do it...

you will
absolutely measure a different SWR (you will measure the reflected
power at point 2), going into 50 Ohms believe it or not.

Yes, you will indeed see varying values of |rho|/SWR on the meter - but
the point you're missing is this: although your imaginary experiment
involves varying two different values called "Z" at the same time, only
*one* of them is having any effect!

The changes you see on the meter are *entirely* due to what you're doing
to the meter itself - changing the Zo for which the meter is calibrated.
Changing the output Z of the PA is having *no* effect on the values of
|rho|. It never does.

You can easily demonstrate this in Real Life (tm) by doing what hams do
every day - adjust the transmitter, which will change its output Z and
its output power; but don't tinker with the calibration Zo of the meter
(which is normally factory-adjusted for life). What you will find is
that the "forward" and "reflected" readings change, but their ratio
|rho| does not (within the limits of accuracy of the meter).

This is effectively the same experiment that I suggested this morning.


Please get
MIMP, it's a free program from Motorola.

Look, I've *written* programs like that, so I know exactly what's going
on under the hood. What we're witnessing here is a demonstration that
such programs can only give answers, not basic understanding. You have
created some construct that seems to explain the program output for you;
but it will not survive real-life experiments.



Or:

PA--+SWR meter1+50 ohm dummy load
(50 Ohm meter)

1 2


In which case, you would NOT measure the reflected power at point 1
(from the discontinuity from the PA to 50 ohms, because this is before
the meter.
This is the case you guys are talking about, WHICH I TOTALLY AGREE
WITH, as the source Z shouldn't affect the SWR meter in this case.


No, the source Z will not affect the |rho|/SWR reading in *any* case.

Go try it - not on the computer, but on the bench.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek

Dr. Slick wrote:
"Ian White, G3SEK" wrote in message
...


You would then have either:

PA--+SWR meter1+50 ohm dummy load
(same Z as PA)

1 2



The changes you see on the meter are *entirely* due to what you're doing
to the meter itself - changing the Zo for which the meter is calibrated.
Changing the output Z of the PA is having *no* effect on the values of
|rho|. It never does.



Nope. you are just changing where the reflections actually happen,
either before or after the SWR meter. In one case you measure it, in
another you don't. As simple as that.

No, it's even simpler. A good SWR meter only *senses* the conditions on
the line - it does not significantly disturb them. You seem to believe
that calibrating an SWR meter for a different Zo "changes where the
reflections actually happen" - it doesn't have to, and if at all
possible it should avoid doing so.

You're now going to claim that an SWR meter contains a section of
transmission line. In such cases you'd be right about the reflections,
but that is a feature of one particular type of meter, and one
particular type of construction - it isn't a fundamental requirement. As
a counter-example, the Bruene bridge meter is not based on a
transmission line, can be made as physically small and non-intrusive as
technology allows, and can be re-calibrated to any Zo by simply tweaking
a capacitor or a pot.

To repeat: a good SWR meter senses only the impedance at its *output*.
And the source impedance has no effect on the readings of |rho| or SWR.



Please get
MIMP, it's a free program from Motorola.

Look, I've *written* programs like that, so I know exactly what's going
on under the hood. What we're witnessing here is a demonstration that
such programs can only give answers, not basic understanding. You have
created some construct that seems to explain the program output for you;
but it will not survive real-life experiments.


Oh REALLY? I'd like to see this program you have written. I doubt
you could give it to us

Oh come on, did you really imagine I was lying?

www.ifwtech.co.uk/g3sek/netcalc.htm
www.marsport.demon.co.uk/scalc.htm
www.marsport.demon.co.uk/sekpmk.htm

These are the ones that come immediately to mind that involve impedance
operations (there are also many others, in other areas). NETcalc is
entirely mine; Scalc is a collaboration; and the programs in the BASIC
software collection are sometimes mine, sometimes improved versions of
very ancient programs by other people. Even when someone else wrote the
original code, I've been 'under the hood' and take technical
responsibility for every line.

NETCALC in particular is a 'Swiss Army Knife' calculator for complex
impedances, which I wrote almost 20 years ago because I needed one.
There still isn't anything else that does quite the same job. It
includes Z0 transformations, and impedance transformation by loss-less
transmission lines, but it doesn't include line losses (never got around
to it) or complex Z0 (something I learned about here, from Reg).
However, if anyone wants to upgrade it...

The BASIC source code for all the programs is there to download (except
Scalc). Some of it isn't elegant, and all of it shows its age, but its
age also ensures that there aren't many technical errors - these
programs have been around for long enough to have been thoroughly
debugged by large numbers of users.

We haven't been maintaining these programs much in recent years, as
others like Reg and VE3ERP are now filling the small BASIC/DOS program
niche very well.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek