¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io
  1. Nanovna-Users
  2. Messages
Expanded Topics Messages
Date
Re: Some basic load measurements

 

There is no second harmonic from the square wave of the Si5351, only the 3rd and 5th harmonic.? Also if you look at the specs of the SA612 it's only up to nominally 500MHz. The latter is pushed way out of spec. If Philip's or whoever changes their chip die, this project is stuffed.Over 200MHz of the Si5351, the instrument's capability seems to rely entirely on the maths routines, FFT or Hilbert.Sent from my Samsung Galaxy smartphone.

-------- Original message --------From: Warren Allgyer <allgyer@...> Date: 08/08/2019 15:00 (GMT+01:00) To: [email protected] Subject: Re: [nanovna-users] Some basic load measurements Elaborating: The "measurement range" quoted by the manufacturer is the range between the displayed carrier level and the displayed noise level. That range deteriorates from 300 - 600 MHz because the lower energy second harmonic is used for those ranges. It further deteriorates for the upper ranges because the third harmonic is used there.The manufacturer does not specify accuracy in measurement and Alan Victor's mathematical observation is borne out by the variance of observed return loss measurements at values lower than -30 dB. Not only are RL measurements below -30 fairly inaccurate, they are also nearly irrelevant for the sort of work this unit was designed for. If you truly need RL measurements below -30 dB then you need to be prepared to spend more than $50.Warren AllgyerWA8TOD
Possible bug?

 

I'm using the latest FW 08_02_19 800_ch version.

I have a possible bug.

Set sweep range:
Start 46 MHz
Stop 54 MHz

Then set Marker | Center.

You'd expect the marker to be a 50 MHz, but instead it's at 46 MHz. If you attempt to move the marker, the sweep range changes.

--

73,
Mike, N1JEZ
"A closed mouth gathers no feet"
Re: Some basic load measurements

 

.... and frequency error on my unit at cold start is 237 Hz at 900 MHz = 0.26 PPM

WA8TOD
Re: Some basic load measurements

 

Elaborating: The "measurement range" quoted by the manufacturer is the range between the displayed carrier level and the displayed noise level. That range deteriorates from 300 - 600 MHz because the lower energy second harmonic is used for those ranges. It further deteriorates for the upper ranges because the third harmonic is used there.

The manufacturer does not specify accuracy in measurement and Alan Victor's mathematical observation is borne out by the variance of observed return loss measurements at values lower than -30 dB.

Not only are RL measurements below -30 fairly inaccurate, they are also nearly irrelevant for the sort of work this unit was designed for. If you truly need RL measurements below -30 dB then you need to be prepared to spend more than $50.

Warren Allgyer
WA8TOD
Re: NanoVna Parts

 

Jim,
Go to your local thrift store or recycling depot (or junk box) and look around the electronics section for portable devices that might use that type of switch. A lot of MP3 headphones use them.

Personally, I'm adding a small button keyboard to mine (see photo). I got the switch flex from a broken Motorola XPR3000 series radio and it fits perfectly after trimming. Call your local 2-way radio dealer service depot and see if they have any broken housings you can have.

I'll post the pin-outs if anyone is interested.

Regards,
Larry

On Thu, Aug 8, 2019 at 05:21 AM, <jimcking@...> wrote:


Anyone know where to get parts for these. I broke the Multifunctional switch
(Jog Switch) on mine and would like to buy a replacement.

Jim K.

buttons.jpg

Re: Some basic load measurements

 

Mine meets that specification.

WA8TOD
Re: Some basic load measurements

 

So is the vendor lying when the manual says:Measurement range: 70dB (50kHz-300MHz), 50dB (300M-600MHz), 40dB(600M-900MHz);Port swr <1.1 Frequency error <0.5ppm?Sent from my Samsung Galaxy smartphone.

-------- Original message --------From: alan victor <avictor73@...> Date: 08/08/2019 05:49 (GMT+01:00) To: [email protected] Subject: Re: [nanovna-users] Some basic load measurements The error mathematics for a one port VNA measurement with SOL is interesting to investigate. Given the assumption that the short and load are handled as ideal with reflection coefficients of -1 and 1, then a finite return loss load is considered. The effect of the finite return loss on the accuracy of the measurement is most pronounced as the DUT has a gamma approaching zero! In? other words as you attempt to measure a device which is MATCHED the measurement error rises significantly. If you expect to measure devices say within a SWR circle of 1.5:1 or greater, you may be working very hard to obtain an improvement that is very minuscule. Most of the devices I believe we generally encounter do not have return losses less than 20 dB. I'll see if I can find or provide the mathematics behind this premise. Alan
Re: NanoVna Parts

 

Is true its handy to have that switch working, but you can still fully
operate the nanovna without it.

Everything from the touchscreen can be done, including drag and drop the
markers...

Cheers!

LL

On Thu, 8 Aug 2019 at 10:21, <jimcking@...> wrote:

Anyone know where to get parts for these. I broke the Multifunctional
switch (Jog Switch) on mine and would like to buy a replacement.

Jim K.
Re: Annotated nanoVNA menu diagram

 

Hi Oristo

This looks great! Really helpful.

I have spotted a couple of differences between it and what I'm seeing with
the 2 trace antenna analyser firmware on my Nano. I believe the 2 trace
version uses a clearer larger font, so the format menu structure is
slightly different.

Here's what I'm seeing on my Nano:

The Trace menu only has Trace 0 and Trace 1.

The Format menu ends with "More" directly after "SWR". Clicking the "More"
button brings up another menu with "Polar" & "Linear" on.

With "Reference position" my default setting is shown as 7 000 for LOGMAG
and changing it to 8 000 moves the trace reference (shown by a small marker
at left of screen) up one grid space, to the top of the screen. With SWR
the trace is off the bottom of the screen when 1:1. Changing the reference
position to 180 brings the trace up onto the bottom of the screen.

Regards

Mike

On Thu, 8 Aug 2019, 00:52 Oristo, <ormpoa@...> wrote:



That URL involves neither JavaScript nor cookies.
It is my first CSS learning exercise and certainly buggy.
For example, branches wrap badly on narrow displays.
It displays OK on an iPad, but hovering requires a mouse..

Hovering a mouse pointer over some text boxes should pop up tooltip hints.
Underlined TEXT have hyperlinks to longer descriptions
in a plain HTML text file on the same website.

TRACE and FORMAT entry annotations are yet to do.
I will also investigate iPad support..
NanoVna Parts

 

Anyone know where to get parts for these. I broke the Multifunctional switch (Jog Switch) on mine and would like to buy a replacement.

Jim K.
Re: Measurement challenge

 

Fahnestock clips take me back to Jr. High School Electric shop, where I got my ham radio license in 1956.
Stuart.
K6YAZ

On Aug 7, 2019, at 6:27 PM, Larry Rothman <ac293@...> wrote:

Fahnestock clips
Re: Battery size

 

Hello All !
with my unit the charging is limited to just under 500ma regardless even tho the charger is I am using is good for twice that much.
So far I am pretty impressed with this little gizmo !
Next step is to compare how close the measurements are with my 8711 & 8712.
got a good deal on my nano, $49 with free shipping !
73 N8AUM
Re: Problem with shorts and load in cal kits

 

Dave,

I solved this to some extent by cutting screwdriver slots in the back of the short and load, allowing the contacts to be held (somewhat) steady while the nut is tightened. It is a bit awkward to co-ordinate everything while tightening. I am thinking of soldering a lever arm to the back of the short to provide a better handle, but that might not be such a good idea for the load with its unknown internal construction.

--John

On Wed, Aug 7, 2019 at 03:23 PM, Dr. David Kirkby from Kirkby Microwave Ltd wrote:


I expect that the short and load in all cal kits included in the NanoVNA
suffer this problem. When you mate two RF connectors you should *always*
rotate the nut, and *not* the whole body of the connector. It is
particularly important with precision connectors, but it is not a good idea
to rotate the body of any connector.

See this link



click on ¡°making connections¡± then go to #4.


1.

CRITICAL: Rotate only the connector nut - NOT THE DEVICE OR CONNECTOR
BODY - until finger-tight, being careful not to cross the threads.
Damage to both connectors will occur if the male center pin is allowed to
rotate in the female contact fingers.

Unfortunately it is impossible to do this on the shorts or loads in the
male cal kit I have. (As I wrote earlier, the opens are best left of)f, but
the problem would not exist with the opens as nothing touches the centre
pin.

Although it¡¯s possible to eliminate the problem with better quality loads,
I see no way to eliminate the problem with the short unless the firmware
was updated to consider the delay of the shorts.

My own companies shorts, opens or loads don't suffer this problem. Whilst
our loads could be used with a NanoVNA, the shorts would produce less
ccurate calibrations due to the firmware limitations of the NanoVNA.

I am unsure if this will make much practical difference on the NanoVNA, but
CrossRF sell similar shorts as in the cal kit I received, *so having a
spare short and a better quality load might be worth owning *
i
Dave.
--
Dr. David Kirkby,
Re: Some basic load measurements

 

The error mathematics for a one port VNA measurement with SOL is interesting to investigate. Given the assumption that the short and load are handled as ideal with reflection coefficients of -1 and 1, then a finite return loss load is considered. The effect of the finite return loss on the accuracy of the measurement is most pronounced as the DUT has a gamma approaching zero! In other words as you attempt to measure a device which is MATCHED the measurement error rises significantly. If you expect to measure devices say within a SWR circle of 1.5:1 or greater, you may be working very hard to obtain an improvement that is very minuscule. Most of the devices I believe we generally encounter do not have return losses less than 20 dB. I'll see if I can find or provide the mathematics behind this premise.

Alan
Re: Some basic load measurements

 

There appears to be TWO different loads provided with the NanoVNA instrument. I reported this in an earlier message. The 50 ohm termination with the white gecko unit was pretty poor and demonstrated a 18 dB return loss at 900 MHz. That has since been degraded to 16 dB after about 3 weeks of using it. When I first measured it NEW it showed a series inductance of approximately 500 pH.

The termination kit 50 ohm with the black VNA unit is considerable better. I measured about 44 dB RTL across the 50 kHz to 900 MHz range. I would consider that quite good. Again as a point of reference, the APC -7 precision connectors provided by Keysight are 53 dB return loss DC-to- 5 GHz. These are the so called sexless connector. Very nice.

I think you should be pleased with a 40 dB + RTL. Far more important, after you find such a wonderful and precise load, take care of it. As soon as you use it and if you allow the center pin to rotate just ONCE, your return loss will be shot! A bit of gold will be scraped off the center pin to rest upon the dielectric. Do that a few times and watch the trail of parasitic L/C begin to build. The return loss is out the window.

Regards, Alan

________________________________
From: [email protected] <[email protected]> on behalf of Peter Gottlieb <hpnpilot@...>
Sent: Thursday, August 8, 2019 1:52 AM
To: [email protected] <[email protected]>
Subject: Re: [nanovna-users] Some basic load measurements

I made some further measurements using the 8753 setup.

I did the OSL cal on one port and looked at both the Smith chart as well as S11.

I didn't bother with the cheap BNC load. I used averaging (16) to get more
stable measurements.

The Smith chart was of little use as each load showed up as a tiny one pixel dot
almost exactly in the middle. However, it does display reactance values.
All the values here are from the 8753.

Load R Z contrib ohms value S11 @ 900 MHz
OSL 50.000 0.0000 0.0000 H -76 dB (noisy)
Nano 49.184 0.4255 76.942 pH -40.673 dB
Narda 50.438 -33.203 m 5.3205 nF -47.356 dB
Tiny 49.389 0.5977 105.58 pH -40.678 dB

Note the resistance values on the 8753 differ from the DC resistance somewhat,
even normalizing to the OSL value. You can clearly see the OSL becomes the
definition of 50 ohms and the S11 is at the analyzer noise floor.

tuckvk3cca pointed out how the 1.02 SWR corresponded to a 40 dB return loss and
he is spot on. The Narda shows the best return loss at 900 MHz of better than
47 dB.

What would be considered a high quality load? The one that comes with the
NanoVNA is not terrible considering the other tiny one I have, which has a NSN
number on it, is very similar. Not that having a NSN number infers anything
spec particular, but at least it will have a minimum set of specs so somebody
thought about it.

I note that the very small SMA terminations are slightly capacitive while the
1.5" long Narda termination (it probably has some power rating) is slightly
inductive. These variations are too small to see on the Smith chart at regular
scale.

Peter


On 8/6/2019 10:43 PM, Peter Gottlieb via Groups.Io wrote:
Resending from website as it didn't seem to go through as a message. Also my pasted table from Excel lost formatting so I tried to fix it to be more readable.


I just did some very simple resistance and SWR measurements using a HP 8753ES with 85046A, resistance was measured using a calibrated Agilent 34401A in 4 wire mode.

I did a very basic one port 3 point cal using a Anritsu OSL which is specified to over 3 GHz.

I took measurements at 900 MHz.

Load R ohms SWR SWR notes
OSL 50.052 1.001 Flat
Cheap BNC 51.104 1.908 Sloping up with freq
Nano load 49.044 1.019 Flat
Narda 12.4 GHz 49.536 1.018 Flat
Tiny SMA 50.787 1.009 Flat


I am guessing there is some significant reactive component in the BNC terminator. All three of the SMA loads showed a flat SWR with frequency so I'm thinking they all have a minimal reactive component.

The difference in resistances while keeping SWR low was a bit of a surprise to me. The load that came with the Nano is over an ohm off of the load I used to calibrate yet the SWR remains at a low 1.019. Why is this? I did the math and surprisingly this is indeed correct, per calculation the SWR should be 1.021 vs my measured 1.019. I'd say this is darn close seeing one measurement is DC resistance and the other is at 900 MHz.

So my conclusion is that SWR is not a sensitive number to see resistance differences.

Once I read some of the references cited I can do some more advanced measurements.

Peter
Re: Measurement challenge

 

Wow, I have not seen those since I was a child, say age 13!

Yes, I suppose they would work fine! What I had in mind and what I am currently having pretty good success with is a Pomoma grabber cable to a BNC connector. I am calibrating in a BNC environment. I adapt from SMA to BNC. Built a BNC short, open and have a commercial hp 50 ohm termination. Not pretty, but 50 kHz to 30 MHz so far reasonably effective. I will try to photo the test set and results.

[cid:42067b24-ec64-4b89-9b13-5fef48015439]

________________________________
From: [email protected] <[email protected]> on behalf of Larry Rothman <ac293@...>
Sent: Thursday, August 8, 2019 1:27 AM
To: [email protected] <[email protected]>
Subject: Re: [nanovna-users] Measurement challenge

Alan,
Alligator clips are so yesterday...
I've got some great Fahnestock clips for you! ;-)

pastedImagefile.png

Test drivers wanted.

 

Anybody wish to volunteer to perform the task of loading and comparing firmware¡¯s and/or versions/release levels?

It would be nice to avoid the judgmental pitting of one against any all others; so I think what is needed and most welcomed by most of us at this stage of our collective level of knowledge and understanding; is a general, but informed, overview with useful comments of how each behaves... not a detailed blow by blow walk through of every operation.

The intent is to establish a sense of stability (bugginess), ergonomic improvements, feature adds/changes, and general user observations to guide users with meaningful and relevant information when assessing if and when a change or upgrade might better serve their own needs.

Not doing so relegates the task to all of us flailing through the exercise on our own. The redundancy of this approach is likely to escalate into multiple divergent and opinionated threads causing the task of making an informed decision unachievable.

That said... If you feel you can take this on before we become overwhelmed with firmwares coming at us from every direction, you will certainly be of high value to the group; both by those that are new to VNA¡¯s, and those who are seasoned, experienced, and most able to contribute in a focused manner to educate and advance our understanding and utility of the NanoVNA in particular, and VNA¡¯s in general.

--
73

Gary, N3GO
Re: Some basic load measurements

 

I made some further measurements using the 8753 setup.

I did the OSL cal on one port and looked at both the Smith chart as well as S11.

I didn't bother with the cheap BNC load.? I used averaging (16) to get more stable measurements.

The Smith chart was of little use as each load showed up as a tiny one pixel dot almost exactly in the middle.? However, it does display reactance values.
All the values here are from the 8753.

Load??????? R??????????????? Z contrib ohms???? value S11 @ 900 MHz
OSL???????? 50.000????????? 0.0000???????????????? 0.0000 H -76 dB (noisy)
Nano????? 49.184????????? 0.4255???????????????? 76.942 pH -40.673 dB
Narda???? 50.438 ? ? ?? -33.203 m? ?????????? 5.3205 nF -47.356 dB
Tiny??????? 49.389 ???????? 0.5977 ???????????????? 105.58 pH -40.678 dB

Note the resistance values on the 8753 differ from the DC resistance somewhat, even normalizing to the OSL value.? You can clearly see the OSL becomes the definition of 50 ohms and the S11 is at the analyzer noise floor.

tuckvk3cca pointed out how the 1.02 SWR corresponded to a 40 dB return loss and he is spot on.? The Narda shows the best return loss at 900 MHz of better than 47 dB.

What would be considered a high quality load?? The one that comes with the NanoVNA is not terrible considering the other tiny one I have, which has a NSN number on it, is very similar.? Not that having a NSN number infers anything spec particular, but at least it will have a minimum set of specs so somebody thought about it.

I note that the very small SMA terminations are slightly capacitive while the 1.5" long Narda termination (it probably has some power rating) is slightly inductive.? These variations are too small to see on the Smith chart at regular scale.

Peter

On 8/6/2019 10:43 PM, Peter Gottlieb via Groups.Io wrote:
Resending from website as it didn't seem to go through as a message. Also my pasted table from Excel lost formatting so I tried to fix it to be more readable.


I just did some very simple resistance and SWR measurements using a HP 8753ES with 85046A, resistance was measured using a calibrated Agilent 34401A in 4 wire mode.

I did a very basic one port 3 point cal using a Anritsu OSL which is specified to over 3 GHz.

I took measurements at 900 MHz.

Load R ohms SWR SWR notes
OSL 50.052 1.001 Flat
Cheap BNC 51.104 1.908 Sloping up with freq
Nano load 49.044 1.019 Flat
Narda 12.4 GHz 49.536 1.018 Flat
Tiny SMA 50.787 1.009 Flat


I am guessing there is some significant reactive component in the BNC terminator. All three of the SMA loads showed a flat SWR with frequency so I'm thinking they all have a minimal reactive component.

The difference in resistances while keeping SWR low was a bit of a surprise to me. The load that came with the Nano is over an ohm off of the load I used to calibrate yet the SWR remains at a low 1.019. Why is this? I did the math and surprisingly this is indeed correct, per calculation the SWR should be 1.021 vs my measured 1.019. I'd say this is darn close seeing one measurement is DC resistance and the other is at 900 MHz.

So my conclusion is that SWR is not a sensitive number to see resistance differences.

Once I read some of the references cited I can do some more advanced measurements.

Peter
Re: Measurement challenge

 

Alan,
Alligator clips are so yesterday...
I've got some great Fahnestock clips for you! ;-)
Re: Problem with shorts and load in cal kits

Dr. David Kirkby from Kirkby Microwave Ltd
 

On Thu, 8 Aug 2019 at 02:08, Gary O'Neil <n3go@...> wrote:

Good observation Dave;

Alan¡¯s suggestion to use SMA Male-Female adapters to protect the
connectors on the NanoVNA would be prudent also.

Both also give us something to scrounge for and haggle a bargain at our
next hamfest. :-)

The care and feeding of RF connectors in general might be a useful topic
for your next YouTube video. As you know improper use of RF connectors can
be equally destructive as using connectors of marginal quality.

--
73

Gary, N3GO

Yes, I had thought about a video on that. I have a video camera, but since
I am a Unix person I had problems getting the data off of the card reader.
I tried a virtual machine, but that didn¡¯t work. I only have the one video
on YouTube, which I took with my mobile whilst doing the task and fighting
hayfever.

Connector savers will not impact the calibration quality, whereas any
attempt to improve the mechanical properties of the short will cause
problems

Dave


--
Dr. David Kirkby,
38741 - 38760 of 39597