开云体育

On Thu, Jul 16, 2020 at 07:50 AM, alan victor wrote:

This measurement technique was addressed many months ago in a prior thread.

I'll note that unwanted shunt-capacitance can result in errors (that increase in frequency) when measuring large impedances using "configuration 3" in the Copper Mountain doc. This same shunt capacitance is why you cannot use S11 to measure large impedances.

Admittedly, the shunt-Cap errors in "configuration 3" measurements are likely to be small. Never the less, the "Y21" method of impedance measurement (see link in first post of this thread) should remove (or significantly reduce) this source of measurement error, should it be a concern.

- Jeff, k6jca

Thanks Darrell I'll take a look. I've been putting off designing one for myself so this may save me some time...

FYI - In the current source code the internal on-chip vbat divider is always enabled. There is an option where the divider is switched on/off before/after the actual vbat measurement via ADC1_17 (based on #define F303_ADC_VREF_ALWAYS_ON).
Hopefully the internal divider is disabled when the MCU is off :)

The required minimal sampling time for the vbat measurement is 2.2us (see the reference manual).
Current ADC setting is
#define ADC_VBAT_SMP_TIME ADC_SMPR_SMP_601P5
that is 614 clocks - around 8us @ 72MHz.

This measurement technique was addressed many months ago in a prior thread.



You can now go to the Copper Mountain web site and read this specific paper that was published in RF and Microwaves earlier in the year.

Alan

These are your only options for emails.

*Individual Messages
You will receive each message in an individual email.

*Full Featured Digest
Group up to 12 messages together as one email instead of being sent individually (rich formatting).

*Plain Digest
Group up to 12 messages together as one email instead of being sent individually (plain formatting).

*Daily Summary
You will receive a daily summary, delivered in the morning, of the group's messages from the previous day.

*Special Notices Only
You will only receive special notices from the moderators.

*No Email
You will receive no emails.

The Vbat, at least what I see in the sources for H4, is being sampled such the Vbat voltage divided by 2 (internal on-chip divider, most probably disconnected when not in use) is converted via the ADC channel 17. There are no external dividers needed.
The rather unusually high noise in the vbat data readings I presented in my other post could be caused by something missing in the H4 code, ie. a small delay should be inserted after activating the divider - before reading the vbat, or something like that..

Just FYI, I designed a protective case for the H4. It is a bit oversized to allow additional SMA connectors to be left on to prevent abuse to the connectors attached directly to the circuit board.
The files are on Thingiverse. I 3D printed mine in PLA but any material should work. There is a bit of room between the connectors at the top to allow a small piece of rubber to be placed if you desire a tighter fit. Look for NanoVNA H4 case or under my designs using Medic433.

Hope you find it useful.

Darrell Carothers

Sent from my over-rated IPhone 7 Plus. Any Mis-spellings or grammar errors are due to my IPhone auto correct feature.

Dear Clifford,

I've studied Dunsmore's publication (thanks for the link!), as I can understand it is more about the way to enhance the U7's input (I'm referring to the 3.4.2) network, namely, the R9~R21.
But as I can see it, there is another cause for worries - the frequency-dependence of SA612's input impedance.
Along with that bridge we also must compensate for SA612's input capacitance. I think it should be possible by connecting an inductor (several nH) in series with the R20 and R21 in CH0 and R25 in CH1.
I do not yet know if this bridge should be used in CH1 too. Most probably - yes, it should be.

Please comment.

Warm regards,
Anton

There have been several threads in the various nanovna forums about the value of applying engineering to these devices.

For the record:
You statements are incorrect, except, perhaps where you imply that you didn't check the manual.
You didn't read the schematics I sent.
You didn't read the accompanying text.
You don't understand how simple the math is.
...and more...
You summarily dismissed something without understanding what you were dismissing. And then got 2 up votes for that. This certainly sets the standard for how we are to interact with developers.

All the NanoVNAs in this class still reflect much creativity and energy, and they are still a very good value for the cost. But they could be so far ahead of where they are now if the developers were more open to sharing ideas and learning from the mistakes of others. I decided to reply here because anyone else reading this thread should be aware the the original post contained information that was completely ignored in the response and the up votes.

--
72/73 de Rich NE1EE
On the banks of the Piscataqua

Nice write-up of the measurement problems. Rings a few bells.

Now with the nano, I measered several resistive loads (47 to 10k) using S11 en S21 in the HF range a few weeks ago. For high impedance values S11 measurements were not even close, as expected.

Here's the measurement of a 10k 5%? 1/4W normal resistor (no smd) from the junkbox using S21.
Reworked calculated series impedance into parallel resistor and capacitor values is what you see in the graph.
I was amazed. My relatively old nano H3.2 + nanoSAVER 0.2.1 do pretty well.

73,
Arie PA3A

Op 16-7-2020 om 01:45 schreef WB2UAQ:

Excellent! Thanks very much, Pete.

73,

Jeff, k6jca

Hi Jeff,
Here's the paper: Showed up below as Challengesandsolutionsforimpedance
When I did the complex algebra it was a good refresher.
73, Pete

On 15.07.20 13:14, MarkZ wrote:

Hi,

I tried this (once) with the latest code and it seemed to work fine.
I also looked through the code and didn't see any reason for what
you're describing. There have been a lot of changes to the export
code recently.

Have you tried with the latest executable?

-Mark

It was a bug releases between v0.3.0 and v0.3.2
fixed in v0.3.3



73
Holger, DG5DBH

This depend from calibration and calibration set, how good it made. And possible you made calibration and use cable, but then made measure not or different.

@DiSlord: while using H4 with v0.8.4.5fw (most stable pre-SDcard version here) I see data outside the Smith Chart (both on the H4's LCD and in nano-saver as well). See below the picture. Is that ok?

I got them a year ago. Links & vendors change...

Just search on Ali...

On Wed, Jul 15, 2020 at 09:08:32AM -0700, Michael Dunn wrote:

I made a simple little video about using the Nano as a spectrum analyzer... Might be of interest to some. Also showing off my RF test bench ;-)

Have you got links for the generator & attenuator?

You should understand NanoVNA can use as SA, but it one point bandwidth (on old firmware as in your variant) ~1kHz so you can see spectre in 100*1kHz = 100kHz span (can be bigger but then you can get more and more error)

If use last firmwares bandwidth can ajusted from 2k to 10Hz (also more fast sweep)

But main limits is points count and max bandwidth.

I made a simple little video about using the Nano as a spectrum analyzer... Might be of interest to some. Also showing off my RF test bench ;-)