开云体育

Well, it's actually an itty-bitty 2-port VNA that, within its specs, provides both reflected and through measurements of RF circuits. So, for they typical ham, mostly antennas, and maybe filters. Considering that most 1-port antenna analyzers are ~$300-900, it's pretty amazing. I found that several evenings reading of VNA basics from the Agilent and Aristu documents to be helpful. After all that, tossing away advanced microwave engineering techniques leaves a decent understanding of using a VNA (along with simulation software like mmana and simsmith) in designing and tuning antennas that are {somewhat longer} than 4cm... ;-)
But seriously... For practical applications, set up the screen to to show SWR, resistance, and reactance, and get out your microscope so you can read the screen, and it will become apparent that dipole length controls reactance (+-jX), and height above ground (for a big horizontal dipole) controls resistance...
Then, since it measures phase, there's the whole issue of "where is the reference plane?" and how to calibrate it at the desired plane. Like, you could calibrate it at the end of a 100 feet of coax, then connect that to an antenna, hoist it 60 feet in the air, and measure the actual feed-point impedance, in its environment, from a convenient distance.
Unfortunately, it isn't likely there will be any detailed written procedures for performing specific tasks, as a VNA is a tool that is adapted to the task at hand. But with only consideration of SWR and R+jX, it's much more useful than a simple SWR meter (or basic antenna analyzer), and it will do much beyond that, if you study into the engineering side of VNA measurement a bit. Collect and read the various PDF documents in the files section, and the many posts about calibrating to a reference plane, and it'll all come together. Then you can set it up to do whatever you want.
As a former industrial instrument technician (having calibrated thousands of things), followed by 4 years of writing procedures, I can tell you it isn't easy! as procedures need to be general enough to cover variables in use-cases, but specific enough to be useful. Indeed, as publisher of a rather large how-to web site, I still find users' sending questions of details not adequately covered, even on long and detailed articles I wrote 20 years ago.
73, --kv5r

#54 : [TheLeastVNA] Code with an Application to [NanoVNA]

Hello,

Allow us, please, to inform you that, as we promised it, we just
uploaded the /F/L/O/S/S [TheLeastVNA] at:



The simple contents of lines from 450 and 780 are the
needed statements for any appropriate language, as we
mentioned at:

#52 : Update : The compact SLO formula for [AnyVNA]
/g/nanovna-users/message/5100
17.October.2019

This code is ready to run under the latest version of
BBCWDEMO.EXE downloadable from:



The example input data and their results are for the sample
frequency of f = ~405 MHz of [NanoVNA], and, as it is easily
verified by inspection, they define the RED points for the
measurements of ( s, l, o, g ) in the figures mentioned at:



#51: the full final report 1 with contiguous full online images is here:
/g/nanovna-users/message/5063

as well as those for Z = ( R , X) mentioned at:

30 : our final report 1 - 6 October 2019
/g/nanovna-users/message/4179

that is at:




Sincerely,

gin&pez@arg

54#

Hi Bryan,
thanks - I figured it was something like that, but it's nice to have it
confirmed :-) With your description of how you did your calibration, I'm
sure more people will be using these fields - so I will prioritise getting
the input validation done :-)

--
Rune / 5Q5R

On Thu, 17 Oct 2019 at 23:17, bryburns via Groups.Io <bryburns=
[email protected]> wrote:

On Wed, Oct 16, 2019 at 07:55 AM, Rune Broberg wrote:

I hope this helps you find the issue that caused the crash.

Rune,

I just figured out what is causing the crash I referenced above. If a
user leaves one of the text boxes in the "Calibration settings" area blank,
no number at all, nanoVNA-Saver will immediately crash resulting in the
error:

File "NanoVNASaver\Calibration.py", line 447, in calculate
ValueError: could not convert string to float:

I think this means that you need to check all of those boxes to make sure
they have a valid number before actually trying to convert them to a
floating point number.

--
Bryan, WA5VAH

On Wed, Oct 16, 2019 at 07:55 AM, Rune Broberg wrote:

I hope this helps you find the issue that caused the crash.

Rune,

I just figured out what is causing the crash I referenced above. If a user leaves one of the text boxes in the "Calibration settings" area blank, no number at all, nanoVNA-Saver will immediately crash resulting in the error:

File "NanoVNASaver\Calibration.py", line 447, in calculate
ValueError: could not convert string to float:

I think this means that you need to check all of those boxes to make sure they have a valid number before actually trying to convert them to a floating point number.

--
Bryan, WA5VAH

Jeff got me going, the driver was not installed. Now to get rid of the Smith chart overlay. BTW, the R and j do come up fine on the Saver, Thanks Rune.

Mel, K6KBE

--
Mel, K6KBE

Hello Mel,
I assume your NanoVNA is plugged in, turned on(!), and that you have tried
pressing the rescan button?

Assuming you are running Windows, you may need a driver installed for your
system to recognise the device. I don't know exactly which one, though, but
I'm sure there's someone on this mailing list who will be able to help you
:-) I think there have been occasional cases of the USB connector being
poorly soldered, and the device not showing up for that reason.

I hope you get it to work!
--
Rune / 5Q5R

On Thu, 17 Oct 2019 at 22:23, Mel Farrer, K6KBE <farrermesa@...>
wrote:

Rune, Good Day,

First time user or want to be, hi. 0.1.3 is downloaded and comes up fine,
but does not present a COM port number? What is the best way to approach
this? Thanks,

Mel, K6KBE

On Thu, Oct 17, 2019 at 1:07 PM R D <rdwd3c@...> wrote:

Rune,

What version of the firmware should I install to use with NanoVNA saver?

Rich, WD3C

Hi Erik -

Using the console command "sample" it is possible to measure the pure
reference, refl and thru signal strength.

Can TAPR VNA save those nanoVNA absolute values to files?

Rune, Good Day,

First time user or want to be, hi. 0.1.3 is downloaded and comes up fine,
but does not present a COM port number? What is the best way to approach
this? Thanks,

Mel, K6KBE

KV5R,
that's great to hear! :D I try what I can to make it scaleable, but I don't
personally have a high-DPI monitor (nor serious vision issues - yet), so I
don't have many opportunities to test it. I hope you continue to enjoy it
:-)

--
Rune / 5Q5R

Hi Rich,
it should work with most versions, but I have good experiences with
edy555's 0.2.3 firmware:



There are instabilities and bugs in firmwares older than maybe
mid-September, some of them causing communications issues, so I would
recommend upgrading if your firmware is older than that.

If you (or others) find a firmware that doesn't work with the app, do let
me know. :-)

--
Rune / 5Q5R

Hi Rune,
I am very happy that your software re-scales everything, even if it does cause a little "jitter" of the graphs' sizes... I use so many software tools that do not re-scale, and everything breaks and is un-readable, when old-timers like me need to use on a 32 inch display with 150% zoom + increase screen fonts just to to see. I love programs that scale everything to the display, zoom, DPI, font points, etc.
Lovin' your software!

Rune,

What version of the firmware should I install to use with NanoVNA saver?

Rich, WD3C

On Thu, Oct 17, 2019 at 03:40 PM, <erik@...> wrote:

LCD refresh is suspended during measurement

OK, great. One less thing to test.

What are your thoughts on the transient response of the 3v regulator data sheet I've attached.

Also, I wonder if you could lower the noise by running from 2 Li cells (7.4v) through a linear 5v regulator.
I assume you've done something like that already.

...Larry

Well, has anyone measured the +5 V rail(s) to see how noisy they are? Once that is done the specs for the parts powered by +5 V would need to be looked at to see how noise is too much.

DaveD

Sent from a small flat thingy

On Thu, Oct 17, 2019 at 11:32 AM, <erik@...> wrote:

Next picture shows what happens when you use a clean 5V supply. Much less
noise in the reference signal.

Erik,

This data seems to suggest that if we had better by-pass capacitors on the 5V supply noise in measurements would be reduced. Would you agree?

--
Bryan, WA5VAH

Hi sala nimi,
any examples please?
Thanks.

Have a fun!

N.V.

LCD refresh is suspended during measurement

On this Ft measurement, I think you should try to measure the full set of S parameters. This simply means turning the test set (whatever that currently is) around end for end and measure S22 and S12. If you were able to measure S21 properly with the 20 dB attenuator on CH0 and not overload CH1, then you may not need any pad on CH1. Your S22 values should give you a mag value close to UNITY. At HF, its going to be pretty high and the reactance will be -X or capacitive. On the mag of S12 in dB for any decent device even a 2N3904, you should see -20 dB or less reverse gain. Then go off and calculate Ft with the full equation.

FUN!

Alan

Rune has the S11 values converted to Z series complex, real and imaginary.

He provides the Q as well and this is the Q series value.

You can transform from series to parallel and he does this correctly by using the formula:

Rp=Rs (Qs^2+1).

The series and parallel network equivalent must have the same Q, hence Qs = Qp.

Then to find the equivalent parallel form as well the series Imaginary form just use the appropriate Q equations, namely:

Qs=Xs/Rs and Qp=Rp/Xp. Hence the X values are found and dependent on the SIGN, choose L or C for the final result.

Alan

Hi Maurizio,
I'm glad you're enjoying the new version :-)

I agree that calibrations should save the start/stop/segments/averaging.
It's a great point. I don't know if I want to overwrite whatever frequency
span the user has set up when they load the calibration, though. It's a
matter of doing what the user expects, and I'm not sure I would expect
loading a calibration to change my sweeping setup? That's my thought for
now, at least.

Showing the marker values on the charts themselves for the values displayed
on the chart is an interesting idea. I will put it on the list of new ideas!

--
Rune / 5Q5R