开云体育

I made a page about this:


Basically, unbalanced bridge sensitivity is non-linear;
for bridges balanced @ 50 Ohms, signal amplitude difference from 0 to 50
is the same as from 50 to 150. Signal change from 500 to 1000 Ohms
is only 10% of change from 0 to 1000 Ohms, so far more sensitive near 50 Ohms,
and "calibrating" with a different LOAD does not help.

In reality, USB does not work that way. Here are "real" supported USB speeds:
USB 1.0/Low-Speed: 1.5 Megabits per second (Mbps)
USB 1.1/Full-Speed: 12 Mbps.
USB 2.0/Hi-Speed: 480 Mbps.
USB 3.0/SuperSpeed: 5 Gbps.
USB 3.1/SuperSpeed: 10 Gbps.

Setting some bps is supported for virtual COM (AKA CDC device) to chips (e.g. FTDi)
at the other end which interface USB to real serial (RS-422 or 232 or whatever)..
There is no such interface at the STM32 chip in nanoVNA,. Speed is determined by
which USB speed gets negotiated at plugin detection time and buffer sizes implemented.

What is the Z range of the NanoVna? I characterized a 500 ohm pot with another Z instrument with the pot set to 500 ohms. At 10 MHz the pot looked something like a resistance 480 ohms with about 80 ohms of reactance. The NanoVna was off by about 20% for R and about 50% off for X. I did some more of this and found the results looked better but only below about 250 or 300 ohms. I know the NanoVNA is not an LCR instrument but I was hoping to see something closer up to about 1K.

I read some discussions about cal kits and all. I do most of my tinkering at HF and use BNC most of the time so I made up a simple BNC short and leave the cable open for an open and attach a 50 ohm load via BNC adpaters for the load. The results are good enough and I do the same for quick measurements at work using the 8753 basically the same way. In any case, none of the connectors I use at home or use at work are type N or APC 3.5 or APC 7 that are supported by real cal kits. Everything is BNC, SMB or SMA or other small connectors in that family.

I've had my NanoVna for about a month now and having a great time doing swept meas'ts at home instead of using my ancient GR bridges. I will not scrap the old bridges because each one has a strength over the others including the NanoVNA.

Hello Rune-

First, thanks for the very useful NanoVNA-Saver application!

Earlier in this thread you mentioned setting the data rate at 115200 bps. This is 2nd hand information- somewhat anecdotal- but apparently plugging multiple devices (mouse, keyboard, etc.) into a USB hub can alter/decrease the data rate for other devices on the hub. I wonder if this can explain any of the cases of reported NanoVNA data loss. I first heard of this over at the SDR-Radio Groups.io list. I realize that 115200 is 'quite slow' by nominal USB standards.

73,

Joe F. KC8RKL

Herb,
You nailed it!? I did the reset and calibration and now the unit works!
I am not so delighted with what I see, as it isn't that close to my static measurements.? However, the unit now behaves as I think it should.? I get believable SWR and Smith presentations.
Thank you so much; I really didn't want to put this thing under the soldering iron.
Bob

Thanks for asking the question. I am familiar with them methods of old. Like sala nimi outlined. The R ch and the A and B ch's are down converted to a low IF and these are applied to the inputs of phase detectors. With the new synthesizer IC's available today the phase of signals can be shifted and the phase can be indirectly determined. I am hoping someone will give us the details!

Hi Bob,
You will find everything in the forum Wiki and Files.
Calibration is probably required, please read the latest manual on how to do it.
There you can get answers to many possible questions.
I'm talking about this: /g/nanovna-users/files/NanoVNA%20User%20Guide-English-reformat-Oct-2-19.pdf

73, Gyula

I haven't looked in detail at the nanoVNA method, but normally you would mix down to a low frequency and do IQ sampling. This is a complex number that can easily be converted into a magnitude and phase.

I understand how a (Wheatstone) bridge gets a voltage or magnitude relative to reactance difference. But how does it get the phase of the that vector.
I think there are some hard ways of getting that phase but surely there are some practical ways of doing that.

Bob,
One other thing to try before you start looking at the hardware.

The first NanoVNA I bought came pre-calibrated and was ready to go out of the box. A second cheaper NanoVNA that I bought came un-calibrated and behaved much as you described. I went through the calibration procedure and saved it to location "0" and the Smith chart display was still pinged to the righthand side regardless of termination. I read a different calibration procedure that said to first select "RESET" under the "CAL" menu and then perform the calibration procedure and save to location "0". After adding that step, the second NanoVNA worked like a champ. You may have already done this but some beginning users skip performing the "RESET" step first.

Herb

Any time you push new F/W to the device it is good practice to re-cal the touchscreen.
Now, with that said, I just updated one of my units and found the touch surface was still calibrated.
So - if you see the touch is off at any time - most likely on the menu selections - do a quick cal.
Regards,Larry

Hmm. Can newly installed firmware detect that it has never been touch screen calibrated and automatically start in TOUCH CAL the first time?

Good day,

When I began this process of working with the NanoVNA, I was not certain how the cal standards were being treated. Since the completed cal showed a perfect open, short and termination, I was pretty certain they were NOT defined as REAL SMA connector devices! The SMA connectors are NOT that perfect!

At the time looking through the code, the best I could determine was they were treated as ideal. The 50 fF C for the open was a new input.

In any case, what I find for RLC measurements is that the actual TEST set is a more significant problem than the calibration pieces. The stray L or C of leads to the DUT must be de-embedded. We could dedicate an entire thread on that alone. However, basically you must take into account the strays of the test set. Long leads and stray C make it more involved. But it is tempting to use that method because it is convenient. Just be prepared to adjust your readings. Again, a measurement of a known component is useful. A so called gold standard that is NOT part of the cal sequence is great for a sanity check.

Alan

Larry, thanks for the encouraging information.? Where do I get the schematics?? What could be wrong with the bridge?? Where do I find it?? Pictures would help.
I don't have much to lose at this point but if I can salvage this it would be great.
Bob

Bob,

It sounds like the bridge circuit on CH0 is bad and the nano is not seeing any reflected energy to measure.Don't throw it away, just do a little troubleshooting - the schematics are available.Make sure there are no shorts going into the DSP from the reflected bridge.
Regards,Larry

Rune, forgive me if you aren't the guy to whom I should ask this.
I purchased a nanoVNA recently and it isn't working as I think it should.? Regardless of terminations I always get a dot at the right edge of the Smith chart.? I always get a flat line for SWR.? However, the frequencies are correct and I can select many menu options.
I don't know if this is a hardware problem (I don't think so).? Is there a way to reset the processor or load some alternate software?? I hate to throw away this lovely little unit, but I may have to get another.
I have calibrated it several times but the problem persists.? At one point I did get a reasonable display but it disappeared and haven't been able to get it back.
For what it's worth, the accuracy of the frequency is amazing.? It's within less than 1 ppm.? For that alone, this unit was worth the price but that's not something I need.? I already have a rubidium standard.
Bob

Hi Mike-

This was a little unclear to me, but interesting. Are you saying that
you used the 51 ohm resistor held by the jaws of the alligator clips for calibration?

Exactly, IMO "calibration" should reflect actual measurement configuration
More specifically, SHORT was by clamping alligators together
and OPEN was by unclamping them.

I want nanoVNA "CAL" to automagically account for alligator clamps
and subsequent readings only about DUTs subsequently clamped.
More "ideal" might be SHORT using a length of wire clamped
between alligators approximately the length of resistor and coil.

Then the data you posted was for that inductor?

Yes, one that is very liable to stray pickup.
The attached image shows nanoVNA display for a frequency scan,
and hopefully visible is an approximating circular segment Smith plot,
as reactances should be.

While component RLC values at intended application wavelength[s] are
obviously of interest,
observing Smith plots over a wider range can reinforce credibility or
reveal anomalies.

Having not been otherwise informed,
nanoVNA reports 50 Ohms for that 51 Ohm resistor;
I can live with 2% errors...

IMO, 5 stored calibrations are none too many.

On Mon, 29 Jul 2019 at 16:59, alan victor <avictor73@...>

So at this point, create the best short, open and 50 ohm load... the

supplied are not too bad. Perhaps the open as suggested in a prior post can
be improved. However, KEY is the VNA is under the impression that these cal
standards are IDEAL.

Perhaps there is a transaction issue, but the source code I looked at
appeared to show that the open was a 50 fF capacitor. Or maybe someone else
modified the source code.


--
Dr. David Kirkby,
Kirkby Microwave Ltd,
drkirkby@...

Telephone 01621-680100./ +44 1621 680100

Registered in England & Wales, company number 08914892.
Registered office:
Stokes Hall Lodge, Burnham Rd, Althorne, Chelmsford, Essex, CM3 6DT, United
Kingdom

RuneSorry for the alert? Win10 was trying to protect me from the download.I deleted? and then reloaded. This time it gave me a functional app.? Now all is well.I will test on thanks you so much for your efforts cheersBill kd5yyk

26 : Correction of Maxima Version and "Unpacked" DERDEI Code

- also @Jose Luu : /g/nanovna-users/message/3278 -

Hello,

Allow us, please, to inform you that, as we already promised at:

21 : The Mathematica Code and the Unpacked DERDEI Code :
/g/nanovna-users/message/3290

we just uploaded the new, "unpacked" version of DERDEI at:


Also allow us, please, to inform, facupov, the interested
Common User that in order to avoid unnecessary
complications, he has to take into account the following:

(0.a/b) As it is detailed in the messages:

Re: [Maxima-discuss] Maxima 5.38.1 64 bit build Windows
From: <petrosez@gm...> - 2016-06-13 11:18:29


Re: [Maxima-discuss] Maxima 5.38.1 64 bit build Windows
From: Gunter K?nigsmann <gunter@pe...> - 2016-06-14 06:44:35


the version 5.38.1 of [maxima] requires a bunch of hard to
follow instructions for installation under wxpp32 or wxpp64.

Therefore, we * e r r o n e o u s l y * suggested, what we had
avoid to say, about this installation at:

2 : yza, Sep 21, 2019 : /g/nanovna-users/message/2802

and thus he should install the last * f u l l y * working 32-bit version
5.36.1 of [maxima] with [sbcl]:


We are terribly sorry for the inconvenience.

Please accept our apologies.

(0.b/b) Note: He may need to also do the following:

|Start|Settings|Control Panel|System|Advanced|Performance|Data
Execution Prevention|(o) Turn on DEP for all programs and
services except those I select:|Add|{X:\}{wxpp32}Program Files\
or{wxpp64}Program Files (x86)\Maxima-sbcl-5.36.1\bin\sbcl.exe|
|Open|

(1) He should unzip [DERDEI20191001.7z] to the default
folder with name: [DERDEI20191001], in the root folder of a drive,
which, if it is different than of our [K:], lets say [X:], then he has
to edit [DERDEI-unpacked_20191001-600-1000.MC] and
in the multi-lines beginning with /*10*/ up to /*15/ to replace
every occurrence of K: by X:

(2) He should skip | ignore any instructions given in the
Documentation:


regarding the file: [functions.txt], because all the contents of this file
are now embedded into: [DERDEI-unpacked_20191001-600-1000.MC]
at the aforementioned lines, and thus he should also ignore
completely the file: [useless-MC-embedded-functions.txt],
which was included into [DERDEI20191001.7z] just for the shake
of completeness

(3) He may execute sequentially, the 15 multi-line commands by
Copy-Paste into [XMAxima] and Enter them one-by-one

(4) He may omit the just informative multi-line commands
beginning with: /*2*/, /*4*/, /*6*/, and /*8*/

(5) If he likes, then in just three 3 steps he can: (I) copy
the whole content of [DERDEI-unpacked_20191001-600-1000.MC],
then: (II) paste them to [maxima], and finally: (III) run it fully.

(6) He may want to change the way in which the six 6 graphics
windows are presented by [gnuplot] and to select between the
"cascading" way, in which each window has to be closed, in order
to appear the next one, or the simultaneous one, by following
the instructions provided by the [Maxima 5.36.0 Manual: 12.
Plotting][12.3 Functions and Variables for Plotting][System
variable: gnuplot_view_args]:

"In Windows versions of Gnuplot older than 4.6.3 the behavior of
"~s -" and "-persist ~s" were the opposite; namely, "-persist
~s" made the plot window and the gnuplot interactive shell
remain, while "~s -" closed the gnuplot shell keeping the plot
window. Therefore, when older gnuplot versions are used in
Windows, it might be necessary to adjust the value of
gnuplot_view_args."

(7) If he is indeed a [mathematica] user, then it is up to him to
transfer -more-or-less "easily"- the graphics commands in the
remaining just four 4 [maxima] multi-line commands beginning
with: /*3*/, /*5*/, /*7*/, and /*9*/, to [mathematica], because:


"Macsyma" [maxima] "was revolutionary in its day, and many later
systems, such as Maple and Mathematica, were inspired by it"
-
at least, if not much more, more than that...

Sincerely,

gin&pez@arg

26