开云体育

I have used several instruments and procedures to measure the isolation in
dB, and impedance in ohms (calculated of course) using K9YC method and S21
on VNAs. They all come terribly close together.
Care must be made in how the test jigs are set up and calibrated. With my
setup, I can easily see return loss of 50-70 dB open circuit. I have
pictures of the various ones I have used if you are interested.

Mel, K6KBE

On Sat, Nov 2, 2019 at 10:32 AM Mel Farrer, K6KBE via Groups.Io <farrermesa=
[email protected]> wrote:

Ed, Check the file section for Balun/choke investigation.

Mel, K6KBE

On Sat, Nov 2, 2019 at 10:23 AM Edward McCann <Ag6cx1@...> wrote:

Who have used the device to measure Impedance to Common Mode Current vs
frequency for bifilar and coax wound ferrite toroid of various materials?

Interested in a step-by-step procedure, which leads tied together,
measurements shield to shield, graphic results, and perhaps pictures of
test jigs, comments on tolerance, etc.

Thanks!!

Ed McCann
AG6CX

On Sat, Nov 2, 2019 at 09:56 AM, QRP RX wrote:

It looks like cheap Chinese SMA-terminators. If so, they have bad performance.

Try to measure it with multimeter, does it have 50.0 ohm?
Most of all it has high error, something like ±5 ohm.

----------------------------------------------------------------------------------------------------------------

As QRP replied, there is a difference between a 50 ohm termination and a good quality 50 ohm calibration load. That is why commercial calibration kits cost so much money. Using an un-characterized termination, your OSL calibration will look correct but you will be referencing all your measurements to however good the termination is. For example if you use a poor termination for performing your OSL calibration and then measure a good quality load. The load will look poor.

Checking the termination with a multimeter only tells you what its DC characteristics are. You need to measure the termination against a known good quality load to verify its characteristics. It might prove worthwhile to construct your own 50 ohm load using the methods presented by other members of this group.

Herb

Impedance to Common Mode Current vs frequency

Some Wiki Application Notes page entries may be helpful
/g/nanovna-users/wiki/Application-Notes

Common-mode chokes by G3TXQ
ferrite bead measurements
test jigs

Ed, Check the file section for Balun/choke investigation.

Mel, K6KBE

You can use NanoVNA-Q firmware, it has scanraw command which don't have such
limit, but it returns raw measurements with no calibration corrections.

For 'scan' command, which can return calibrated values,
NanoVNA-Q firmware currently ignores the [points] value, always returning 101,
but 'frequencies' returns the number of points requested.

Using 'scan' command, e.g. in edy555 firmware,
a shell script can combine multiple scans, e.g.:

ch scan
usage: scan {start(Hz)} {stop(Hz)} [points]
ch pause
ch scan 10000 100000 90
ch data 0 > data0.txt
ch data 1 > data1.txt
ch frequencies > freq.txt
ch scan 100100 200100 90
ch data 0 >> data0.txt
ch data 1 >> data1.txt
ch frequencies >> freq.txt
...

Who have used the device to measure Impedance to Common Mode Current vs frequency for bifilar and coax wound ferrite toroid of various materials?

Interested in a step-by-step procedure, which leads tied together, measurements shield to shield, graphic results, and perhaps pictures of test jigs, comments on tolerance, etc.

Thanks!!

Ed McCann
AG6CX

On 23/10/2019 20:21, John Nightingale via Groups.Io wrote:
--------------snip------------------

A little note for new Linux users who find that Rune's application

is not seeing the NanoVNA when it is physically connected: root's
privileges will be required in running the application so that the
application is permitted to make a connection to the NanoVNA. If the
issue is that the application has not been permitted to open a port,
there will be a report to that effect in the console display.

Please don't encourage people to run Linux GUI applications as root. It's a very bad idea.

It's likely that the regular user is not in the dialout group which is easily rectified by entering the following command in a terminal:

sudo usermod -G dialout $USER

You will then need to log out and back in once, after which nanovna-saver should always be run as a regular user.

Cheers,
Barry
G4MKT

Kurt wanted to see some pictures so here they are.

The first picture is the dB logmag of S11 of a 30cm semi rigid cable (medium) and a 1 meter RG58 cable (long), both Open and Short at the end, both from 0.5 MHz till 900MHz
When looking at the polar picture it is obvious the wobble of the logmag at lower frequencies is caused by a shift of the center of the open/short rotation as the lines do not cross (see third polar plot of the long cable 0-300MHz)
In the second picture a 106fF Open C0 correction is applied (as automatically calculated by Octave). This shifts the rotation just enough to remove the crossing of the Open/Short logmag
When I let Octave calculate a cable impedance correction (assuming a small mismatch between the cable impedance and Z0) I get the fourth picture shows almost no wobble. The impedance correction is a shift in the order of +0.025 + 0.5i ohm

For know I am running out of ideas on how to proceed as I do have no good model for the semi rigid cable
Anyone that knows how cables behave in a VNA?

Erik, PD0EK

On Sat, Nov 2, 2019 at 06:50 PM, @arnoh wrote:

at RFPARTS to buy for 3,95$

It looks like cheap Chinese SMA-terminators. If so, they have bad performance.
Try to measure it with multimeter, does it have 50.0 ohm?
Most of all it has high error, something like ±5 ohm.

Mine failed also - a piece of junk. Bought my own online.

On Sat, Nov 2, 2019 at 06:47 PM, hwalker wrote:

On the other hand, its probably unclear to a lot of users how to generate the
equal number of calibration points to correct the raw measurements. Any
suggestions for doing so?

You can do scanraw for O,S,L cal-kit loads and use these results to calculate calibration corrections.
After that just apply these corrections to a scanraw result.

at RFPARTS to buy for 3,95$

Had recently gained bad experiece with cheap SMA-load baught via ebay (aro 1$ a piece), none of them had 50ohms,
so I discarded them, contacted the ebay dealer for replacement, got it, but they were no good either

On Sat, Nov 2, 2019 at 09:28 AM, QRP RX wrote:

You can use NanoVNA-Q firmware, it has scanraw command which don't have such
limit, but it returns raw measurements with no calibration corrections. You're
needs to perform your own calibration procedure.

----------------------------------------------------------------------------------------------------------------------

Its great that the scanraw command allows you to return more than 101 measurement points. On the other hand, its probably unclear to a lot of users how to generate the equal number of calibration points to correct the raw measurements. Any suggestions for doing so?

- Herb

On Sat, Nov 2, 2019 at 06:39 PM, Oristo wrote:

FWIW, by setting 'threshold 330000000',
my nanoVNA gets crazy
but no PLL icon

With thershold more than 300 MHz si5351 will use wrong frequency.
Hardware works ok, so there is no PLL icon, but it uses wrong frequencies, so it won't works properly.
Some invalid thereshold values you may hung your si5351 chip and it will needs power off in order to restore normal functions.

But I think, most of all, this issue is related to some garbage in the settings (bad configuration settings).

Another way to fix bad configuration settings is to cleanup all memory before firmware update. It can be done with ERASE.DFU file.

With NanoVNA-Q v0.4.3 it should not happens, because it has a different magic signature for settings and if there is a bad configuration settings, it should be discarded.

invalid frequencies

FWIW, by setting 'threshold 330000000',
my nanoVNA gets crazy CH0 and CH1 logmag and Smith @ 983140600 Hz
but no PLL icon, perhaps because power-on default scan does not go over 100MHz.

On Sat, Nov 2, 2019 at 06:34 PM, hwalker wrote:

Does that mean that if threshold frequency is set to 250 MHz then instead of 3
x 300 MHz = 900 MHz the operating limit will be 3 x 250 MHz = 750 MHz?

yes

On Sat, Nov 2, 2019 at 09:19 AM, QRP RX wrote:

If it doesn't help and PLL icon is still there, it means that your si5351 chip
cannot work at 300 MHz.
You can fix it by reduce threshold frequency below 300 MHz and use saveconfig
command to save that change.
It means that working range will be less than normal si5351.

------------------------------------------------------------------------------------------

Does that mean that if threshold frequency is set to 250 MHz then instead of 3 x 300 MHz = 900 MHz the operating limit will be 3 x 250 MHz = 750 MHz?

- Herb

On Sat, Nov 2, 2019 at 05:34 PM, Maurizio IZ1MDJ wrote:

I think is not possibile to have more than 101 points.

Yes, this is not possible to get more that 101 points with data command. Because firmware has fixed 101 points and you cannot change that.

You can use NanoVNA-Q firmware, it has scanraw command which don't have such limit, but it returns raw measurements with no calibration corrections. You're needs to perform your own calibration procedure.

50 ohm cal load failed

SMA connectors are fragile, and provided load spins in nanoVNA female socket.
Some folks reduce CH0 SMA connector wear by adding male-female adapter.

For frequent recalibration and BNC-equipped DUTs,
I and others buy BNC cal kits, which are IMO more durable and convenient, e.g.: