开云体育

On Sat, Jul 6, 2019 at 03:07 PM, alan victor wrote:

Here is a 1 dB SMA pad terminated in a short. The cal was done at the end of a
SMA connector saver. The 1 dB pad connected to the saver and the pad shorted
with their short SMA cal. The Gamma should be 0.87, the marker read out is ~
0.83-to-0.85 for the most part till 900 MHz where the Gamma drops to ~ 0.8.

Here is a 1 dB SMA pad terminated in a short. The cal was done at the end of a SMA connector saver. The 1 dB pad connected to the saver and the pad shorted with their short SMA cal. The Gamma should be 0.87, the marker read out is ~ 0.83-to-0.85 for the most part till 900 MHz where the Gamma drops to ~ 0.8.

Alan

________________________________
From: [email protected] <[email protected]> on behalf of alan victor <avictor73@...>
Sent: Saturday, July 6, 2019 9:34 PM
To: [email protected]
Subject: Re: [nanovna-users] Messaging (over USB-C) between nanovna and PC client - documented (or could be)?

Ernst,

I have done something similar. I will see if I can gather up the worse case that I have considered. A calibrated 1 dB SMA pad terminated in a short. Hence, a 2 dB return loss, a reflection coefficient of ~ 0.8

________________________________
From: [email protected] <[email protected]> on behalf of dk1vi <dk1vi@...>
Sent: Saturday, July 6, 2019 8:35 PM
To: [email protected]
Subject: Re: [nanovna-users] Messaging (over USB-C) between nanovna and PC client - documented (or could be)?

I have commercial mismatches specified up to 4GHz with N-cnnectors (m). After calibration with the supplied SMA calset I mesured the mismatches connected via an SMA (m) to N(f) adapter.
The result was very convincing, see enclosed plot.
Ernst

Ernst,

I have done something similar. I will see if I can gather up the worse case that I have considered. A calibrated 1 dB SMA pad terminated in a short. Hence, a 2 dB return loss, a reflection coefficient of ~ 0.8

________________________________
From: [email protected] <[email protected]> on behalf of dk1vi <dk1vi@...>
Sent: Saturday, July 6, 2019 8:35 PM
To: [email protected]
Subject: Re: [nanovna-users] Messaging (over USB-C) between nanovna and PC client - documented (or could be)?

I have commercial mismatches specified up to 4GHz with N-cnnectors (m). After calibration with the supplied SMA calset I mesured the mismatches connected via an SMA (m) to N(f) adapter.
The result was very convincing, see enclosed plot.
Ernst

I have commercial mismatches specified up to 4GHz with N-cnnectors (m). After calibration with the supplied SMA calset I mesured the mismatches connected via an SMA (m) to N(f) adapter.
The result was very convincing, see enclosed plot.
Ernst

Hi Tim

My main purpose in experimenting with the NanoVNA was to see if it could make a good impedance analyzer.

Unfortunately, results so far indicate that it does not perform well in this respect. As you probably know, a VNA can be used to measure impedance using the reflecting bridge to measure S11. Comparing results from the Nano to a professional impedance analyzer with the same SOL calibration shows substantial differences.

It did perform well measuring small inductors (1 uH ) from 50 KHz to 1 MHz. Above that frequency errors increased substantially.

There are several possible reasons for the disappointing results. We can discuss this more if you are interested.

But, because of this performance I cannot, in good conscience, release a program that misleads users by producing inferior results. If you or someone else pursues this path, please make sure you have a good impedance analyzer and some good reference components for comparison.

Regards,

George

Hi Larry,

I also did not understand Hugen comment.

However, this is a guess based on what I see when the PC SW connects and pulls measured data from the nanoVNA instrument.

This only occurs when you request the SW to read that measured data. After extracting correct measured data, it appears that the PC SW alters the cal state of the nanoVNA. A reset which is a soft touch screen icon on the nanoVNA may put back the prior corrected cal state. Or if you were to do a cal from the PC SW, there is an icon for that on the PC SW screen, the same reset on the touch screen is required.

I have used the PC SW to capture real time tuning of a matching network. The PC will track the nanoVNA point for point, say a log mag plot of S11 or a S11 Smith plot and this occurs very rapid and with no errors. So the SW does several functions very well with no hiccups. But it is not perfect. And glitches between the SW and HW do occur.

Alan

________________________________
From: [email protected] <[email protected]> on behalf of larry@... <larry@...>
Sent: Saturday, July 6, 2019 1:15 PM
To: [email protected]
Subject: Re: [nanovna-users] Strange peaks in S11 with load

What do you mean by reset cal data.

Do you mean turn cal correction off before trying to collect new cal data.

What key strocks are you talking about.

Do you believe the unit is using modified readings instead of raw data to calculate correction array

larry

-----Original Message-----
From: [email protected] <[email protected]> On Behalf Of hugen@...
Sent: Friday, July 5, 2019 8:52 PM
To: [email protected]
Subject: Re: [nanovna-users] Strange peaks in S11 with load

It is necessary to reset the calibration data before carrying out the calibration.

FWIW this finally shipped, So less then worried now :) Here is to hoping its not a bad clone once it gets here!

Thanks

--

73,
Mike, N1JEZ
"A closed mouth gathers no feet"

Can someone re-post the link or show what message the google drive link is?
Thanks
Frank

What do you mean by reset cal data.

Do you mean turn cal correction off before trying to collect new cal data.

What key strocks are you talking about.

Do you believe the unit is using modified readings instead of raw data to calculate correction array

larry

It is necessary to reset the calibration data before carrying out the calibration.

Hi George Could i perhaps loop in with you ... looking to build something better for Linux seems silly to double up the work if i can 'adjust' your code ... also thinking an Android app might be a nice idea for portable use (take measurements and save them to your cloud etc in the field?).

Cheers

Tim

hugen,
from outside your device looks same as the "worse clone". So only a look inside will tell if you got a good one.
So I enclosed a picture of my device, hoping it is a good clone.
Ernst

Hi Ernst. Not sure I understand your question. And the correction switch, is not clear to me either just yet. However, see my message 83. That is the sequence I have followed from the touch screen and the results after cal are reasonable. I have NOT done a cal from the PC software at this time.

Alan

________________________________
From: [email protected] <[email protected]> on behalf of dk1vi <dk1vi@...>
Sent: Friday, July 5, 2019 5:29 PM
To: [email protected]
Subject: Re: [nanovna-users] Strange peaks in S11 with load

Still confused. When performing calibration from within NanoVNA V 1.03 all is well.
Maybe I did something wrong when I calibrate from the touch screen.
So what does the correction switch do, which is not available on the PC software?
Ernst

Still confused. When performing calibration from within NanoVNA V 1.03 all is well.
Maybe I did something wrong when I calibrate from the touch screen.
So what does the correction switch do, which is not available on the PC software?
Ernst

Hi Ernst,

When the calibration is complete does the side screen show all the elements of the cal. I assume you are doing an ISOL cal as well as a THRU cal despite perhaps not interested in a S21 measure.

Alan



________________________________
From: [email protected] <[email protected]> on behalf of dk1vi <dk1vi@...>
Sent: Friday, July 5, 2019 3:57 PM
To: [email protected]
Subject: Re: [nanovna-users] Strange peaks in S11 with load

[Edited Message Follows]
[Reason: Correction of calibration range]

Flashed firmware nanoVNA_800_ch_20190705.dfu. Unit still initialises from 50k to 900M.
I performed calibration from 50k to 800M, but behavior is strange: there is always an offset of abt -24dB on displayed mag S11 during calibration process..
Now the plots are smooth without spikes but calibration does not work as expected.
Short gives 0dB but open shows -9dB. With "Correction" switched, open and short both have S11 -24dB.
Ernst

Flashed firmware nanoVNA_800_ch_20190705.dfu. Unit still initialises from 50k to 900M.
I performed calibration from 50k to 800M, but behavior is strange: there is always an offset of abt -24dB on displayed mag S11 during calibration process..
Now the plots are smooth without spikes but calibration does not work as expected.
I made calibration with Software on PC was connected.
Recalibrating with PC disconneted give good results. However with load, mag S11 is very rugged. When using "Correction" line gets smooth, but with short and open, there is -25dB offset. Strange.
Ernst

Just found new firmeware dated 20190705 on the well known google drive.
Ernst

Folks,
There appear to be two locations on GitHub for the nanoVNA firmware:
original:
hugen:

Both appear to be old and haven't been updated in some time.

It would be great if Hugen would provide a URL link to the opensource code he is currently maintaining for this device or at least make some commits to the repository.

Thanks,
Larry