开云体育

@Erik, PD0EK - 6 November 2019:
/g/nanovna-users/message/6539

Dear Erik, This is not a stupid question. Best regards, gin&pez@arg

PS More on that by Gary, N3GO - 6 November 2019:
/g/nanovna-users/message/6566

On Tue, Nov 5, 2019 at 10:47 AM, Oristo wrote:

Adding a ferrite core improved results with USB connection:

Unfortunately ferrite don't helps, the problem with IP5303, the result is worse when external voltage is applied. It doesn't matter if voltage is pure clean (from the battery) or has some kind of PC noise.

But you can wait when charge LED will stops to blink (charge end), after that it works better with USB connection. Not as good as with no cable, but close to that.

On Wed, Nov 6, 2019 at 09:25 PM, Jim Allyn - N7JA wrote:

Windows Vista is out of support and no longer updated, so, if you continue to
use it, at a minimum you should keep it off the internet.? I have updated
many older computers running Vista, XP (and even earlier versions) to Linux.

I'm using Windows XP, it works great. Linux with X11 is much slower and requires more RAM.
Internet on Windows XP also works ok. Some issues with SLL, but there is exists patch from MS, which allows to fix SSL issues.
Firefox works ok. :)

Nothing in TDR apart from the reflection at the end
I'm using a 1000 point calibration exactly matching the measurement points
--
Erik, PD0EK

When I tried to install Python on a machine that didn't have it,
but I didn't know it because it was missing a piece of Ms VC ++.
In my opinion, if your basic operating system is OK,
you can install MS VC ++ (multipack 2005-2017) on it without error,
and the nanoVNA-Saver that uses a lot of graphics will work.

For Vista, v2015-14.0.24215 is wanted according to this:

I chose a resistor to avoid derailing discussions about quality of cables.

Before finally dismissing them, does TDR show some reflection at 3x cable length?
If yes, then calibration does not fully compensate CH0 mismatch.

Also, calibration for 200-900MHz stimulus will make a little better use of 101 points of correction.

I successfully started nanoVNA-Saver when I installed the SP1 service pack and Ms VC ++ multipack on windows7-32 bit.
When I tried to install Python on a machine that didn't have it, but I didn't know it because it was missing a piece of Ms VC ++.
In my opinion, if your basic operating system is OK, you can install MS VC ++ (multipack 2005-2017) on it without error, and the nanoVNA-Saver that uses a lot of graphics will work.
Without him, not even Python.

73, Gyula HA3HZ

Measuring the same 25ohm resistor connected in the same way on my three VNA's . All three have been calibrated with the same calibration kit and show perfect calibration. (SOL at -1,+1,0 for all frequencies)
1: nanoVNA at reduced power: The reduction of power above 300MHz has made the measurement less erratic but there is still an unexplained drop in magnitude of S11 with frequency
2: ownVNA (similar design as nanoVNA but dead bug style build). Some erratic behavior but the drop with frequency seen in the nanoVNA is absent
3: 3GHzVNA same measurement method as nanoVNA but no harmonics mode . Most stable measurement with some remaining cable length pattern (explainable from the design) and erratic patterns around 160MHz,350MHz and 650MHz

Any suggestions on what can cause the drop in S11 in the nanoVNA that does NOT disappear with calibration?
A non-linearity of some component like the mixers, but how to test apart from reducing power?
I will measure if the drop is the same for different resistors

--
Erik, PD0EK

Here is definitive evidence there is something to be improved in the nanoVNA design when measuring at higher frequencies (above 300MHz)
DUT is a 25 ohm resistor close to the reference plane with a small amount of parallel capacitance
I chose a resistor to avoid derailing discussions about quality of cables.
Two measurements are done:

1: Default nanoVNA settings (first picture)
2: Reduced power in harmonics mode (second picture)

The reduced power measurement makes more sense


One ghost found, now use other easily modelled and understood components (resistors, inductors, capacitors) to find more.

--
Erik, PD0EK

Coaxial cable loss frequency depend. Higher frequency higher loss. But swr 3 with low or high impedance not same loss. Simple model not support this. Advanced model good. Now attenuation / frequency graph not like a line. Small like sine waves on line. Not zero lenght not ideal oslt also like a lossy coax cable. Many cheap chinese connector not tetlon dielectric. Loss not small.

Simple (dB/m) and advanced (k parameters) coaxial cable modell exist.

Does not fully understand this forum. Cable loss compensated? I see big errors from cable loss. Lossy coaxial line not circle on smith diagram. It s a spiral. I see electrical lenght compensated, but loss not.

Thank you,
Just flashing using latest AA version from
2 traces is enaough for me.

Port 1 mismatch is already corrected for by the usual SOL 3 term calibration. The only things left that it *doesn't* correct for are linearity error (e.g. signal source's output power changing depending on DUT impedance in a nonlinear way) and side channels. In this case we have two nearly identical impedances (the open circuit standard vs the cable's impedance which crosses the open circuit point) that measure wildly differently, so the most likely explanation is side channels (other things that are different about the two DUTs other than the impedance at the measured frequency), which one of them is the DUT's impedance at f/3. The solution is, of course, don't use harmonic mode ;)

No problem, we're all here to help each other.?



On Wed, 6 Nov 2019 at 7:05 PM, QRP RX<qrp.ddc@...> wrote: Larry Rothman, sorry. I opened groups.io, found question which I seen in email and wrote answer. Just didn't noticed that I replied on your answer :)

Just FYI and to keep with the original post.
I have manged to extract multiple "projects" from simply cloning the original Github repository.
So part of the goal is done.
Now I have to figure out which of the multitude of C projects is the chief.
I can add a new file, using plain "project" perspective - which is a step up.
There seems to be some test / emulation code projects too.

As far as RFI from Bluetooth hardware - will cross that river when I get there,
Still looking for software partner(s).

I tried to create V 0.1.5 for my mac, OS version is Catalina.
I got an error message trying to install scipy…

I think scipy was new for v0.1.5?

Has anyone else encountered this?

Dana VE3DS

Larry Rothman, sorry. I opened groups.io, found question which I seen in email and wrote answer. Just didn't noticed that I replied on your answer :)

QRP, that is not my quote, below - that is from the other member whom I already explained what you just did as well....?


On Wed, 6 Nov 2019 at 6:54 PM, QRP RX<qrp.ddc@...> wrote: On Thu, Nov? 7, 2019 at 01:41 AM, Larry Rothman wrote:

I don't know where I can enter vbat_offset.
Can anyone "lead my hand"?

just connect to your NanoVNA with PuTTY and use command "vbat_offset"

This command will change battery offset until power off.
In order to save current battery offset permanently, use command "saveconfig"

On Thu, Nov 7, 2019 at 01:41 AM, Larry Rothman wrote:

I don't know where I can enter vbat_offset.
Can anyone "lead my hand"?

just connect to your NanoVNA with PuTTY and use command "vbat_offset"

This command will change battery offset until power off.
In order to save current battery offset permanently, use command "saveconfig"