开云体育

When developing my nanovna mini-tools, I am unclear about the "Touchstone" file format
regarding the storage of Z-parameters for a one-port (nanovna_snp.py). So far I follow rev
1.1, it says on page 7:

In a .S1P file it is not an impedance that is described but S11

# HZ S RI R 50
6800000 0.4178235144152254 -0.6694732263322638

For each line you find in order:
Frequency in Hz
the real part of S11
the imaginary part of S11

If you want the impedances you have to recalculate them from S11
--
F1AMM (Fran?ois)

-----Message d'origine-----

De la part de Ho-Ro
mardi 26 juillet 2022 09:47

I attach simple Excel file for the same example.

It uses some of the in-built functions for dealing with complex numbers.

A complete list of available functions can be found at

Kind regards

Ed, G8FAX

Hi,



When developing my nanovna mini-tools, I am unclear about the "Touchstone" file format regarding the storage of Z-parameters for a one-port (nanovna_snp.py). So far I follow rev 1.1, it says on page 7:

Example 2:
!1-port Z-parameter file, multiple frequency points
# MHz Z MA R 75
!freq magZ11 angZ11
100 0.99 -4
200 0.80 -22
300 0.707 -45
400 0.40 -62
500 0.01 -89
Note that in the above example Z11 (the input impedance) is normalized
to 75 ohms, as given by the reference impedance (R 75) in the option
line.

Unfortunately rev. 2.0 is not totally clear about the Z format, on the one hand it is not normalized (page 7):

For Version 2.0 files, the reference resistance defines the system reference for the S-parameter data if the
[Reference] keyword is not present. Network data for G-, H-, Y- and Z-parameters in Version 2.0 files is
not normalized. Therefore, the reference resistance and [Reference] keyword have no impact on G-, H-, Y-
, or Z-parameter data in Version 2.0 files. S-parameters are, by definition, normalized with respect to the
reference impedance(s) and in this respect there is no difference between the treatment of S-parameters in
Version 1.0 and Version 2.0 files.

On the other hand, normalization of Z is mentioned in this option line example on the next page 8:

Frequency in Hz, Z-parameters in magnitude-angle format, normalized to 10 ohms:
# Hz Z MA R 10

But I suspect that this information is not correct, because further back in the document on page 14 rev 1.1 and rev 2.0 are contrasted for Z-parameters:

Example 9 (Version 1.0):
!1-port Z-parameter file, multiple frequency points
# MHz Z MA R 75
!freq magZ11 angZ11
100 0.99 -4
200 0.80 -22
300 0.707 -45
400 0.40 -62
500 0.01 -89
Note that, in the above example, Z11 is normalized to 75 ohms, as given by the reference impedance (R 75)
in the option line.

Example 10 (Version 2.0):
!1-port Z-parameter file, multiple frequency points
[Version] 2.0
# MHz Z MA
[Number of Ports] 1
[Number of Frequencies] 5
[Reference] 20.0
[Network Data]
!freq magZ11 angZ11
100 74.25 -4
200 60 -22
300 53.025 -45
400 30 -62
500 0.75 -89

This example duplicates the data in Example 9, using Version 2.0 syntax. Note that normalization does not
apply.

Question to the specialists, should one stick with rev 1.1, as does nanovna-saver, or is rev 2.0 more widely used.

A quick test with scikit-rf shows that it writes also in 1.1 format, but creates only S-Parameter files, even if fed with Z-parameter:

!Created with skrf...
# Hz S RI R 50.0
!freq ReS11 ImS11
50000.0 0.98... -0.01...
...

Martin

Thanks Ed, AG6CX,

An interesting read.

Kind regards

Ed, G8FAX

Hi Francois,

Here is a worked example in a simple way avoiding manipulations with the complex conjugate.

Based on the example at bottom of page 3 of the document.

I will start off with Zn = 2.33 + j3.33 in formula S11 = GAMMA = (Zn-1) / (Zn +1)

Step 1 Let Fr = Zn-1 and Gr = zn + 1 so S11 = F / G

Step 2. Substitute for Zn in Fr, so Fr = 2.33 + j3.33 - 1 = 1.33 + J3.33

Step 3. Convert Fr to polar form in the traditional way. I am sure you know how to do this. I do not cover this here and used a calculator

The result is Fp = 3.59 <68.23

Step 4. Substitute Zn in Gr, so Gr = 3.33 + J3.33

Step 5. Convert Gr to polar form to give Gp = 4.71 <45

Step 6. Divide in polar form S11 = Fp / Gp = 3.59 / 4.71 < 68.23 -45 = 0.7622 <23.23

Step 7. Convert polar to rectangular form in the usual way or use, like me here a calculator

S11 = 0.7 + j0.3

I hope that helps

Kind regards

Ed

Study Pilloud. You have no excuse! It’s written in French

I had already browsed this website

I did not find the text relevant and, as a result, I did not trust what I was reading there. It lacks, as often, many developments to understand. Example :

Back to top 3
It does not indicate that, in the Smith chart, the rectangular coordinates are S11. It is however the most convenient way to position an impedance on the chart, especially by software.

Bottom of page 3
We find the formula
(S11 = Zn-1)/(Zn+1) = 0.7 + j0.3
It's not educational at all. How do we go from one to the other.

Page 4 Suitably adapted line (Ligne convenablement adaptée)
It is a totally false idea, peddled from site to site and from microphone to earphone, to say that the generator must have an internal impedance of 50 ? for the line to be adapted. The adaptation is independent of the impedance of the source. Fortunately otherwise our Tx would have a deplorable performance.

Thank you for your explanations which have the merit of being much clearer
--
Fran?ois

-----Message d'origine-----

De la part de AG6CX
lundi 25 juillet 2022 22:03

Hi Doug
The 234/freq is a average, not an exact value.
The real value varies with the diameter of the wire element.
Much like the lad who puts 7 gal of petrol in his car, drives 243 miles,and says he got 34.71428532 miles per gallon.?

It's a good starting point, but you do not need to measure to fractions of an inch.
Cheers? Kent G8EMY/W5

On 7/25/22 1:38 PM, W0LEV wrote:

Also remember that a 1/4-wavelength radiator over a "perfect" image plane
(a.k.a. "ground" plane) is not 50-ohms but more like 36 ohms. So, even at
resonance, defined by +jX = - jX, the SWR will not be 1:1.

What's a lot of fun is to make a ground plane antenna, and then hook up a VNA, and watch how it changes as you bend things.

People spend a lot of time trying to build matching networks, when often, all you need is to make the angles different.

Copper foil tape on paper is another one.

Or, get a bunch of cheap telescoping whips and you can make a two (or three) element yagi and fool with different spacings, lengths, etc.

Dave - W?LEV
On Mon, Jul 25, 2022 at 7:54 PM Greg Giglio <coffeeguy2@...>
wrote:

Generally, look for where the most pronounced ‘dip’ is relative to the
frequency you’re tuning for. It should be close to what you’re looking
for. If the lowest SWR is at a lower frequency than expected, shorten the
wire. If it’s at a higher frequency, lengthen the wire.

73 de KN7GIG

"Silence is golden. Duct tape is silver."

On Jul 25, 2022, at 14:33, DOUGLAS SEWELL via groups.io <douglassewell=

[email protected]> wrote:

?Using formula 234/freq for a 1/4 wave vertical with 4 radials I was

building gave me a wire length of 16 feet 6 and 1/8 inches. Nano vna gave
swr of 2.3 (too high for my qrp station) so tried to find out whether to
shorten or lengthen wire. How can I use the nano vna to help with the
direction to go. eventually lengthened by 6 inches which brought swr down
to 1.3.

As a newbe to short wave radio tx/rx and to nano vna can anyone help

with an easy to follow method to use my nano vna for wire adjustment
direction. Thanks but difficult just now after suffering 2 strokes so hence
the need for help here.

Thanks and 73 to all who read this.
Doug MM7DSA

To those interested in Woodward's work, I found the 1953 paper that was referenced in the recent link.
Sent FYI.

Ed McCann
AG6CX
Sausalito

The ini file for NanoVNA Saver is stored in this location in Windows.

C:\Users\Your user name here\AppData\Roaming\NanoVNASaver\NanoVNASaver.ini

Roger

For those wishing to wind their own baluns or even use/test commercial ones, the following may be of interest:

Performance of balun with/without center-tap

Fundamentals of rf transformers, includes various tests of such

Test configurations and circuits for testing baluns

Kind regards

Ed G8FAX

Hello Jim,

I already have small cores, but I am not convinced it was the right
material. but I will also try to rewind 1015 transformer, another DIY
project;-)
I like to test always few 'roads'...
By the way, it needs to be a unbalanced to balanced transformer, so on the
balanced side a centre tap.


Op ma 25 jul. 2022 om 21:31 schreef Jim Lux <jimlux@...>:

You'll not be rewinding that transformer. The assembly is 6 mm (1/4") on
a side, and it is "not user serviceable", in that the winding is probably
bonded to the core, the core is bonded into the tiny plastic package, etc.
And then, you don't know how many turns on each side. or the wire gauge
(other than "very small" - it's probably AWG40 or smaller)

The pictures of the transfomers look doable, I am sure I will need my
magnifying glass (say 4mm core).

I saw a few RF transformers of MCL:

Might but these also. Beside my own DIY.

You'd have a tough time beating the <$4 each that MiniCircuits wants for

something like an ADT1-6T.

Agree. But still, I want to have some fun also;-)

All the best,

Victor

Also remember that a 1/4-wavelength radiator over a "perfect" image plane
(a.k.a. "ground" plane) is not 50-ohms but more like 36 ohms. So, even at
resonance, defined by +jX = - jX, the SWR will not be 1:1.

Dave - W?LEV

On Mon, Jul 25, 2022 at 7:54 PM Greg Giglio <coffeeguy2@...>
wrote:

Generally, look for where the most pronounced ‘dip’ is relative to the
frequency you’re tuning for. It should be close to what you’re looking
for. If the lowest SWR is at a lower frequency than expected, shorten the
wire. If it’s at a higher frequency, lengthen the wire.

73 de KN7GIG

"Silence is golden. Duct tape is silver."

On Jul 25, 2022, at 14:33, DOUGLAS SEWELL via groups.io <douglassewell=

[email protected]> wrote:

?Using formula 234/freq for a 1/4 wave vertical with 4 radials I was

building gave me a wire length of 16 feet 6 and 1/8 inches. Nano vna gave
swr of 2.3 (too high for my qrp station) so tried to find out whether to
shorten or lengthen wire. How can I use the nano vna to help with the
direction to go. eventually lengthened by 6 inches which brought swr down
to 1.3.

As a newbe to short wave radio tx/rx and to nano vna can anyone help

with an easy to follow method to use my nano vna for wire adjustment
direction. Thanks but difficult just now after suffering 2 strokes so hence
the need for help here.

Thanks and 73 to all who read this.
Doug MM7DSA

--
*Dave - W?LEV*
*Just Let Darwin Work*


--
Dave - W?LEV

Don't look at just SWR!! Look at the complex impedance R +/- jX. Look at
the sign of the complex term, the +/-jX. If it is negative, it is
capacitive and needs lengthening. If it is positive, it is inductive and
requires shortening.

Dave - W?LEV

On Mon, Jul 25, 2022 at 6:33 PM DOUGLAS SEWELL via groups.io <douglassewell=
[email protected]> wrote:

Using formula 234/freq for a 1/4 wave vertical with 4 radials I was
building gave me a wire length of 16 feet 6 and 1/8 inches. Nano vna gave
swr of 2.3 (too high for my qrp station) so tried to find out whether to
shorten or lengthen wire. How can I use the nano vna to help with the
direction to go. eventually lengthened by 6 inches which brought swr down
to 1.3.
As a newbe to short wave radio tx/rx and to nano vna can anyone help with
an easy to follow method to use my nano vna for wire adjustment direction.
Thanks but difficult just now after suffering 2 strokes so hence the need
for help here.
Thanks and 73 to all who read this.
Doug MM7DSA

--
*Dave - W?LEV*
*Just Let Darwin Work*


--
Dave - W?LEV

On 7/25/22 11:23 AM, DOUGLAS SEWELL via groups.io wrote:

Using formula 234/freq for a 1/4 wave vertical with 4 radials I was building gave me a wire length of 16 feet 6 and 1/8 inches. Nano vna gave swr of 2.3 (too high for my qrp station) so tried to find out whether to shorten or lengthen wire. How can I use the nano vna to help with the direction to go. eventually lengthened by 6 inches which brought swr down to 1.3.
As a newbe to short wave radio tx/rx and to nano vna can anyone help with an easy to follow method to use my nano vna for wire adjustment direction. Thanks but difficult just now after suffering 2 strokes so hence the need for help here.
Thanks and 73 to all who read this.

Run the SWR vs frequency plot. Shortening moves the frequency ( at the minimum ) higher.

For a ground plane vertical, all the elements have an effect on resonant frequency and the depth of the SWR.

The length of the vertical and the radials all mostly affect the reactive part; where the reactance crosses zero is defined as resonance.

The "droop angle" mostly sets the "resistance" at resonance. A dipole (1/4 wave radials essentially straight down) has a resistance of around 70 ohms. If the radials point straight out (90 degrees) then the resistance is closer to a monopole over ground (35 ohms). Of course, changing the angle also changes the reactance.


This is where a VNA is fun.. you can build a model, hook it up, and change them both to see what happens.

I've had a lot of fun just building a dipole out of a stiff wire (e.g. AWG 12 bare wire) for something like 200 MHz. You bring the feed line up a support (like a stick or mic stand) with a RF toroid or two as a balun.

What I usually do is cut the wire long, then fold it back to shorten it.
But you can also do things like just make 90 degree bends (like a Moxon).

Bonjour Francois:

To your first point, your expanded expression in the seventh paragraph for gamma in terms of A,B, and Zo is mathematically correct.

But I will say it is not a form I have ever seen used, and while correct, I find, for my purposes, not particularly useful.

As far as the basic question on the fifth paragraph of your post, the question indeed is answered in the attachment. (For your convenience, I attach the link to Pilloud’s description, written en Francaise)

Typical steps in finding S11 from impedances follow below. But note finding S11 is typically one step along the way towards calculating VSWR and Reflection Coefficient.

General steps are:

1. Measure Zs = Rs + Zs.
Zs will be a vector in the form a + jb

2. Calculate ABS Zs = I Zs |
In terms of SQRT (Rs*Rs + Xs*Xs)

3. Calculate Impedance Phase
Angle Zs = arctan (Xs/Rs)

4. Note Zo

5. Calculate Gamma = (Zs-Zo)/(Zs +Zo)
in vector form using the Excel complex division feature, or by hand is so inclined.

6. The resultant vector will indeed also be in the form Gamma= a + jb

7. Gamma vector will have magnitude rho and Gamma angle, often called Theta, when converted to polar form using the same technique you used with impedance, above.

9. Note Gamma and S11 are the same, and are sometimes referred to as the Reflection Coefficient, a complex quantity, usually discussed in terms of rho and Theta.

8. Once all this is in hand, you can calculate
VSWR = ( 1+ Mag Gamma)/(1-Mag Gamma)

8 As noted above, Mag Gamma is referred to as rho, so VSWR = (1+rho)/(1-rho)

9. You can calculate Return Loss (RL)
using RL = -20 log(rho) in terms of dB.

You will find in this site various Excel-based programs that will allow you to derive Z from S11 and S21, as well as starting with Zs in the form Rs and Xs to derive Gamma and all the other parameters of interest.

My iPhone probably has means to express Gamma, rho, Theta in symbols but I’ve not yet found the appropriate app.

Study Pilloud. You have no excuse! It’s written in French





Ed McCann
AG6CX

Jim is 100% right!

Not worth "redoing" it, possibly not even possible. Not to say that when you are done you need to "weld" your wire to the soldering pad (SMD) in a way that will not get disconnected when you attempt to solder to your PCB :)

In the real spirit of DIY, I'd get a bit larger ferrite cores, a it thicker wire (AWG36 probably) and experiment with the ferrite material used, number of turns, geometry of windings, measure across the frequency range, and after a week of having fun go to Mouser and by a factory made one :)

Generally, look for where the most pronounced ‘dip’ is relative to the frequency you’re tuning for. It should be close to what you’re looking for. If the lowest SWR is at a lower frequency than expected, shorten the wire. If it’s at a higher frequency, lengthen the wire.

73 de KN7GIG

"Silence is golden. Duct tape is silver."

You'll not be rewinding that transformer. The assembly is 6 mm (1/4") on a side, and it is "not user serviceable", in that the winding is probably bonded to the core, the core is bonded into the tiny plastic package, etc. And then, you don't know how many turns on each side. or the wire gauge (other than "very small" - it's probably AWG40 or smaller)

I'd just buy a transformer with what you need. Or use it at 1.5:1 - if you're hooking it up to a VNA, it doesn't really matter, because you would be calibrating out the turns ratio in any case.

or buy some cores in a more practical size and wind your own. You'll need to make sure you get the right mix, and deal with the whole wire size, number of turns, how to put the windings on so the capacitance is well controlled. You'd have a tough time beating the <$4 each that MiniCircuits wants for something like an ADT1-6T.

Farnell (in UK), Newark, Digikey, Mouser all carry various mfrs of transformers - Murata, Johansen, Coilcraft, and Minicircuits are probably the most likely.

I added a python program to my toolbox that retrieves the S-parameter from NanoVNA-H and stores them either as S-parameter or Z-parameter in touchstone RI (real/imag) format. It would be easy to create also MA (magnitude/angle) format for Z-parameter.



Developed and tested on Linux, but should also work on Windows when you adapt the serial port name.

usage: nanovna_snp.py [-h] [-o [FILE]] [-p [PORT]] [--s2p] [-z]

optional arguments:
-h, --help show this help message and exit
-o [FILE], --out [FILE]
write output to FILE, default = sys.stdout
-p [PORT], --port [PORT]
connect to serial port PORT, default = /dev/ttyACM0
--s2p fetch also s21 parameter in addition to s11
-z, --format_z fetch s11 and calculate R +jX