开云体育

I removed the comma from the end of the message link below. It works now.?



On Sun, 16 Feb 2020 at 4:39 PM, hwalker<herbwalker2476@...> wrote: ? I agree with Kurt that the NanoVNA's inaccuracies above 300 ohms are not that bad.? I had the same frustrations with an HP Agilent 8753C Network Analyzer with 85046A S-Parameter Test Set. Once you get very far from 50 ohms its accuracy also suffers.? An HP applications engineer I contacted explained that's pretty much the nature of the beast for any low impedance 50 ohm VNA, and there's no reason to expect any more of the NanoVNA.? As Kurt said, the test fixture also plays an important part in measurement accuracy, especially above 1 MHz.? Flying leads test fixtures basically only work for quick checks in the kHz region.

? Erik wrote a Matlab script, /g/nanovna-users/message/7685?to?use an external bridge to calibrate the NanoVNA at other than 50 ohms, but I never saw any feedback from users who tried using it at 300, 1000, or higher impedances.? There were actually about 34 messages in the "Using an external RF bridge with NanoVNA " topic and I don't remember a definitive test procedure write-up coming out of it.

- Herb

I agree with Kurt that the NanoVNA's inaccuracies above 300 ohms are not that bad. I had the same frustrations with an HP Agilent 8753C Network Analyzer with 85046A S-Parameter Test Set. Once you get very far from 50 ohms its accuracy also suffers. An HP applications engineer I contacted explained that's pretty much the nature of the beast for any low impedance 50 ohm VNA, and there's no reason to expect any more of the NanoVNA. As Kurt said, the test fixture also plays an important part in measurement accuracy, especially above 1 MHz. Flying leads test fixtures basically only work for quick checks in the kHz region.

Erik wrote a Matlab script, /g/nanovna-users/message/7685, to use an external bridge to calibrate the NanoVNA at other than 50 ohms, but I never saw any feedback from users who tried using it at 300, 1000, or higher impedances. There were actually about 34 messages in the "Using an external RF bridge with NanoVNA " topic and I don't remember a definitive test procedure write-up coming out of it.

- Herb

Since my phone has a MicroUSB,. I just ordered this one. If it works, then
I may try to find a way to update the wiki (?) so others can have one to go
to directly.


USB C to Micro USB OTG Cable, CableCreation 0.65 ft Type C Braided Cord,
480Mbps Compatible with MacBook (Pro), Galaxy S8, S9, S10, Pixel 3 XL, 2
XL, Android Devices, 0.2M/ Space Gray

Hi WB2UAQ
I do not agree the NanoVNA is that bad but rather think you measurement setup is bad and for tht ametter the calibration. You might forget that if you read the R+jX with the NanoVNA the resistive part falls quickly if you testjig is not calibrated out of the "formula" Just try to measure the Rp instead of Rs then you see why. It is a quite commom mistake forget that if you have a small shunt C og a small series L the across frequency range the Rs drops and Rp stays more stable at the real value. NanoVNA-saver has good facilities for both proper calibration and presentation of Rs+jXs as well as Rp and Xp
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af WB2UAQ
Sendt: 16. februar 2020 15:52
Til: [email protected]
Emne: Re: [nanovna-users] nanoVNA Maximum Impedance Measurement (Folded Dipole example)

I found that above 300 ohms or so the nano has trouble with accuracy. I used a GR-1606-A to measure an impedance and then compared that to the nano and as the Z got higher the accuracy diminished. Oristo explained the reason for this to me and it is because of the limitation of the resistive bridge. Even HP/Agilent/Keysight VNA's accuracy degrades at high impedances but not nearly as bad as the nano. They use directional couplers instead of resistive bridges (Wheatstone basically). One could use the nano with a directional coupler in S21 mode and do an open/short/load cal on it. The S21 mode returns the amplitude and phase angle of the reflected signal relative to the incident signal and from it you can calculate Z and all the other parameters. I have not worked on this yet but have used the HP8721A coupler in S21 as a great scalar return loss measuring system with about 50-60 dB of range and with excellent directivity.

In regard to measuring a typical axial lead resistor, many have shunt capacitances of 0.5 pF and up so this becomes significant at RF. Some are even spiral-wound.

Yes, but I used this one

As it also can power the attached device (NanoVNA)

There are a few reasons why the NanoVNA is not that accurate for high
impedance.

1. The resistive bridge doesn't do a great job at exploiting the dynamic
range of the SA612 mixer. The maximum input voltage you can input without
significant nonlinear mixing products is about 10 mV. The minimum is
somewhere between 1 and 10 uV. So maybe 60-70 dB of dynamic range is
available if its mapped properly to the available load impedances.

2. To compensate for this lack of dynamic range, the SA602 is overdriven
so that its higher order mixing products are significant. The LO input is
way overdriven (it should be 300 mV p-p max) and the RF input should kept
be less than 10 mV. Because the bridge is designed to provide more of a
signal when the load is not balanced to 50 ohms, this nonlinearity becomes
more pronounced for loads that deviate from 50 ohms.

3. All of the gain is provided by the codec and it relies on the codec for
the anti-aliasing filter as well. So to realize the 16-bit dynamic range
of the codec, the codec needs to resolve the original RF signal, if limited
to 10 mV, a resolution of 150 nV before amplification. The full-scale p-p
signal input to the codec is 1 V rms according to the datasheet, so
something like 40 dB total gain (25 dB after including 15 dB of gain of
SA612) is needed to fully utilize the dynamic range of the ADC.

4. The acquisition is not synchronized to the IF signal so that for long
integrations phase error in the IF demodulation can accumulate, which
limits the amount of additional dynamic range one can obtain from long
integrations.

The bridge produces more signal for more imbalance but it doesn't
necessarily produce more accurate results. A bridge that measures the
current in the load is specialized more to measure high impedance loads but
must be very linear to be able to get accurate impedances.

On Sun, Feb 16, 2020 at 08:14 AM, Kayak wrote:

just ordered a "ugreen" otg cable from amazon. It is micro-usb to usb-a female, into that I guess I plug the usb C cable from the nano
=====================================================

Hopefully, you've read some of the past message threads regarding the Android app. The apk installation from the Playstore is experimental and may not detect the NanoVNA even if you have the proper cable. The Web client at usually connects and works, but you need to have internet access to use it.

- Herb

will this cable, if used with the usb-c cable that came with my
nano-vna-h, allow OTG communication?

interesting and something to keep a look out for. I ordered one from
amazon and I can plug the phone's regular micro-usb cable into it.
come to think, maybe that wont work... jeez...

as to your RFI issues, I wonder if you coiled the cable small, and
wrapped in alum foil would limit that?

On Sat, Feb 15, 2020 at 11:16 PM David Fragale via Groups.Io
<DWFComm@...> wrote:

Out of sheer surprise, I found out that the cable- I had bought thinking I was going to use it to transfer some pix from one smartphone to another - winds up being perfect for a patch between my "H" and My LG V30. It's a SCOSCHE Strikeline USB-c "charge,sync and power" CC23WT, I found at Staples for about $20 I think. By the way, I had been dealing with some phantom instabilities with the cable provided, which got a lot worse with the longer cable. Often, the way I dress or orientate the cable between the "H" and Smart Phone would cause some really strange response curves . This at least doubled when I put the longer cable in play. This was much more touchy as I went up in frequency and with the nearness of metal hardware (antennas tubing and supports and such). I seem to have found that the "H" is sensitive to the ground field as it effects some apparent feedback in to the device. In pure antenna theory thought mode I put two of the largest ferrite cable half shells sets around each end of the USB-c jumper..and the vast incidents of strange readings went away. I still have several peeves with traces disappearing and test plot points placement under preset ones, but at least the sudden radical frequency response slides have gone away. I suspect that this is just the curse of a tiny hypersensitive box in a non lab environment , but this cure is pretty simple .
DWF, KA4CNP, Seminole,Fla

In a message dated 2/15/2020 4:49:19 PM Eastern Standard Time, w9td@... writes:

You need a “ on the go” cable (OTG)

Correct me if I am wrong, Once calibrated with those 259's than it should't matter ???

thank you just ordered a "ugreen" otg cable from amazon. It is
micro-usb to usb-a female, into that I guess I plug the usb C cable
from the nano

Thanks to Larry for getting me the AA flash. I was pleasantly surprised to see how well this little box works, The readings after calibration were dead on the money.

Very glad I did not toss it in the garbage. Anyone having problems just hang in there, it is totally worth the effort!

73,

Bill K4FX

I found that above 300 ohms or so the nano has trouble with accuracy. I used a GR-1606-A to measure an impedance and then compared that to the nano and as the Z got higher the accuracy diminished. Oristo explained the reason for this to me and it is because of the limitation of the resistive bridge. Even HP/Agilent/Keysight VNA's accuracy degrades at high impedances but not nearly as bad as the nano. They use directional couplers instead of resistive bridges (Wheatstone basically). One could use the nano with a directional coupler in S21 mode and do an open/short/load cal on it. The S21 mode returns the amplitude and phase angle of the reflected signal relative to the incident signal and from it you can calculate Z and all the other parameters. I have not worked on this yet but have used the HP8721A coupler in S21 as a great scalar return loss measuring system with about 50-60 dB of range and with excellent directivity.

In regard to measuring a typical axial lead resistor, many have shunt capacitances of 0.5 pF and up so this becomes significant at RF. Some are even spiral-wound.

Out of sheer surprise, I found out that the cable- I had bought thinking I was going to use it to transfer some pix from one smartphone to another - winds up being perfect for a patch between my "H" and My LG V30.? ?It's a SCOSCHE Strikeline USB-c "charge,sync and power" CC23WT, I found at Staples for about $20 I think.? ?By the way, I had been dealing with some phantom instabilities with the cable provided, which got a lot worse with the longer cable. Often, the way I dress or orientate the cable between the "H" and Smart Phone would cause some really strange response curves . This at least doubled when I put the longer cable in play. This was much more touchy as I went up in frequency and with the nearness of metal hardware (antennas? tubing and supports and such).? ? I seem to have found that the "H" is sensitive to the ground field as it effects some apparent feedback in to the device. In pure antenna theory thought mode? I put two of the largest ferrite cable half shells sets around each end of the USB-c jumper..and the vast incidents of strange readings went away.? ? I still have several peeves with traces disappearing and test plot points placement under preset ones, but at least the sudden radical frequency response slides have gone away.? ? I suspect that this is just the curse of a tiny hypersensitive box in a non lab environment , but this cure is pretty simple .
? ? ?DWF, KA4CNP, Seminole,Fla

Larry,
I greatly appreciate that advice because, believe it or not, I could not find that info on the the Wiki or on Hugen79's Github pages (which are Greek to me). I didn't find any indication of the different uses for the .bin .hex or .dfu files.

I don't know sw development; neither skilled or interested in building apps or fw - which appears to be much of the wiki content. I just want the most capable tool that presumably the most recent fw would give.

So your comment, while off-the-cuff, was more helpful that you realize; thanks again!

The best specific guidance I found was here:

-- Regards,
I_B_Nbridgema!

You need a “ on the go” cable (OTG)

hi all, I couldnt find in the wiki. I have a Micro-USB on my phone so
bought a Micro-USB to USB C cable from amazon and the app will not see
my nanoVNA-H (from noo).

The cable did say for data so it's not just a charge only cable.
amazon.com/gp/product/B0744BKDRD

Any advice appreciated to get this going thank you.

On Sat, Feb 15, 2020 at 08:06 AM, Russ@va3rr wrote:

So, I received notice that my H4 was shipped by China post yesterday. Do I need M3 sized washers to address the display issue?
====================================================================================
Russ,
hugen posted the following about a week ago:

" I'm very sorry for this problem, and I will fix the plastic case as soon as the Chinese factory resumes normal work."

So the issue may already have been addressed in the H4 you are receiving. If not, then in most cases you do not get the "stops at boot screen" symptom but may notice the following.

If a screen menu is brought up and pressure is applied to the right side of the case, random menu choices occur.

In that case, installing the washer should correct the problem. On my H4 you have to deliberately squeeze the right side pretty hard to notice the problem, so I haven't bothered installing the washer.

- Herb

So, I received notice that my H4 was shipped by China post yesterday. Do I need M3 sized washers to address the display issue?
tia de va3rr

his main email inbox is kept 'clean'

Using a gmail filter, all nanovna-users messages are routed to a folder,
then I mute many topics, but see any new topics (at least once before muting).