开云体育

I'd like to suggest is collecting calibration data and doing statistical analyses.

With calibration turned off, collect SOLT for 101 points over the 70 KHz to 900 MHz range using the console interface and post them to the list as ASCII text: frequency, magnitude and phase.

Having a large amount of such measurements would be immensely beneficial to developing better FW and calibration for the devices. And as OEM QA.

To be useful FW and HW version and thru cable metadata is needed.

The nanoVNA is an order of magnitude reduction in cost for a VNA. It's not the equal of a VNWA or similar, but for many it is good enough and cheap enough for price not to be an issue.

Let's help make designing and building QRP gear something a student can do at home. I endured many failures as a grad student for lack of access to T&M gear. I was not an EE student, so I couldn't use the EE lab facilities. The geosciences department had lots of instruments, but none for doing RF measurements. So all I had was what I could afford to buy, which was *very* little. Getting a 5 MHz recurrent sweep Heathkit IO-18 at a yard sale for $75 was a major event in my life.

We've been handed the ball. Let's run with it!

Have Fun!
Reg

You are right. We should definitely start something. May be a portal to upload all this data and analyze it.

--
Salil Tembe
(Nuclearrambo.com)

In console mode you can use data to dump the calibration data
Would it be useful to collect this as people will have this data available
usage: data [2-6]
where
2 /* error term directivity */
3 /* error term source match */
4 /* error term reflection tracking */
5 /* error term transmission tracking */
6 /* error term isolation */

And maybe also ask edy555 what data would help him to further improve?

The SOLT results tell us what the calibration terms *should* be.

short and open are reflection coefficients of +1 and -1.

load should be S11 of zero

thru should be a linear phase delay for S21.

with uncorrected measurements we can look for systematic errors in the design and fabrication.

On Tue, 10 Sep 2019 at 14:00, Reginald Beardsley via Groups.Io <pulaskite=
[email protected]> wrote:

The SOLT results tell us what the calibration terms *should* be.

short and open are reflection coefficients of +1 and -1.

load should be S11 of zero

thru should be a linear phase delay for S21.

with uncorrected measurements we can look for systematic errors in the
design and fabrication.

It is technically *impossible* to achieve those S-parameters in certain
connectors. Even ignoring the inductance of shorts, and fringing
capacitance of opens, you can not get around the fact that the physical
construction of some connectors just prevents this. The easiest to see is
the female N connector. Just look down one and see the centre conductor.
Look about a 1/5th inch (5 mm) lower and you will see a ridge where the
outer outer conductors mate. That is the reference plane. So the inner
connector sticks up several mm above the reference plane, making it
impossible to achieve these parameters.

It’s just technically impossible to do this

Dave.

--

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

The entire point of calibration is to compensate for the non-ideal nature of physical realizations.

Any deviation from the ideal case is an error term which needs to be accounted for and removed by deconvolution , "deembedding" or "calibration" in EE speak.

In addition to the linear phase terms caused by non-coincident reference planes, there are frequency dependent phase terms caused by stray capacitance and inductance. If the stray reactance is not located at the reference plane, it will have a linear component which results from the travel time difference. This is an easily solved math problem.

Because it is easier to pick the correct travel time on a symmetric waveform, in seismic work it is routine to record an impulse response for the anti-alias filters and ADC chain and then phase and amplitude correct the data to broaden the BW and remove the filter phase delay. At least among the more meticulous seismic processors. The dominant effect is the two zero and single pole filter produced by reflections between the free surface and the water bottom. So often the instrument corrections get lumped in with those. Generally by people who are also not doing the water column deconvolution step correctly.

I'm still adjusting to the difference between the pure real elastic case and the complex electromagnetic case, but it's not that big a step. The wave equation is still the wave equation. And I was doing light in anisotropic media for several years before I ever got involved with elastic wave propagation.

I'm proposing collecting raw, unaltered data and using a statistical analysis of that to determine what the calibration routines should do. I am well aware that this has been done in various forms. But I always feel more comfortable if I go back to first principles when solving a problem. There are very often approximations made which are not described well and may be significant in certain cases.

On reflection it's probably because if someone came to me with a question, it was because the usual solution was not working.

Have Fun!
Reg

Reg,
Here are some ideas and suggestions if you decide to do this.

I suggest you setup a github repository and add the data to it. It is free,
and won't be going away anytime soon, easy enough once you get the hang of
it. Browsing and downloading files is very easy for anyone else.

Use s1p format for the one port files and s2p for the 2 port. It already
contains freq, mag, phase.

Take sweeps of the open/short and load at the same time as sweeping the
unknown and save all of the raw measurements. This provides your 'control',
and if at some point you get your OSL standards swept by someone with a
proper VNA cal kit then _all_ of the measurements you have taken can be
error corrected later.

You will find some 'raw data' sweeps in my repo already, including the
attenuator from my cal kit which has a reference measurement from Dr Kirkby.


Explanation of the directories:

data/ideals = s1p files for perfect cal standards. For accurate results
replace these with actual measurements of the standards measured with a
calibrated VNA - to allow for the imperfect standards.
data/measurements = sweeps of the standards and attenuators on the 8753 and
nanoVNA - all are pre-calibration 'raw data' except for the files which say
'calibrated;
data/output = error corrected results of the attenuator - i.e. post
calibration
data/reference = reference measurement of the attenuator by Dr Kirkby.

(Note, the error-corrected output files do not match the reference
measurement in this case simply because I used perfect calibration
standards in this test. I was doing a comparison between the 8753 and the
Nano doesn't yet support properly characterised cal standards)

Roger


On Wed, 11 Sep 2019 at 05:39, Reginald Beardsley via Groups.Io <pulaskite=
[email protected]> wrote:

The entire point of calibration is to compensate for the non-ideal nature
of physical realizations.

Any deviation from the ideal case is an error term which needs to be
accounted for and removed by deconvolution , "deembedding" or
"calibration" in EE speak.

In addition to the linear phase terms caused by non-coincident reference
planes, there are frequency dependent phase terms caused by stray
capacitance and inductance. If the stray reactance is not located at the
reference plane, it will have a linear component which results from the
travel time difference. This is an easily solved math problem.

Because it is easier to pick the correct travel time on a symmetric
waveform, in seismic work it is routine to record an impulse response for
the anti-alias filters and ADC chain and then phase and amplitude correct
the data to broaden the BW and remove the filter phase delay. At least
among the more meticulous seismic processors. The dominant effect is the
two zero and single pole filter produced by reflections between the free
surface and the water bottom. So often the instrument corrections get
lumped in with those. Generally by people who are also not doing the water
column deconvolution step correctly.

I'm still adjusting to the difference between the pure real elastic case
and the complex electromagnetic case, but it's not that big a step. The
wave equation is still the wave equation. And I was doing light in
anisotropic media for several years before I ever got involved with elastic
wave propagation.

I'm proposing collecting raw, unaltered data and using a statistical
analysis of that to determine what the calibration routines should do. I
am well aware that this has been done in various forms. But I always feel
more comfortable if I go back to first principles when solving a problem.
There are very often approximations made which are not described well and
may be significant in certain cases.

On reflection it's probably because if someone came to me with a question,
it was because the usual solution was not working.

Have Fun!
Reg

On Tue, 10 Sep 2019 at 18:39, Reginald Beardsley via Groups.Io <pulaskite=
[email protected]> wrote:

I'm proposing collecting raw, unaltered data and using a statistical
analysis of that to determine what the calibration routines should do. I
am well aware that this has been done in various forms. But I always feel
more comfortable if I go back to first principles when solving a problem.
There are very often approximations made which are not described well and
may be significant in certain cases.

A few things you might want to be aware of Reg. Perhaps I have
misunderstood what

*1) Mating similar connectors *
There are several connector types which will mate, although they are not
identical

1) SMA, 3.5 mm and 2.92 mm (K connector)
2) 2.4 mm & 1.85 mm

When different types are mated, there’s a discontinuity which has been
analysed and can be corrected for. There’s a paper that does this for

1) 3.5 mm and 2.92 mm (K connector)
2) 2.4 mm & 1.85 mm

The author(s) looked at the correction for SMA, but concluded that the
variability on the semi-precision SMA connector was too large to make it
worthwhile.


*2) Polynomial fits*

The usual polynomial is certainly not based on any accurate physical model.
However, for almost any high-end commercial calibration kit below about 50
GHz, that is what you will be given. You *might* be given measured data
too.

Given the accuracy of the polynomial fit method *far* exceeds that
warranted for an SMA connector, I can’t help feeling that coming up with a
physical model based an an SMA connector is of no practical use whatsoever.

Perhaps I have misunderstood your intentions.

Have Fun!
Reg

Dave

--

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

I'm suggesting something very basic. Accumulate enough data to make a meaningful statement of the mean and variance of S11 and S21 using the parts supplied with the unit by doing uncalibrated SOLT measurements with two open measurements. The bare SMA connector and the supplied open. Both ports for 5 cases, and if possible, a 6th case with both ports terminated with a 50 ohm load. No unknowns. Just poor quality "knowns".

Many of the people who buy these do not, and will not, have access to anything better. An Amphenol cal kit from SDR-Kits is roughly 1/2 the cost of the nanoVNA. The Rosenberger cal kit is 2x the price of a nanoVNA.

This is hobby science. The purpose is to provide a statistical estimate of measurement uncertainty. Fundamentally calibration consists of assuming that any deviation of the measured values from the theoretical values is device and fixture error. In this case we also have significant errors in the calibration standards.

Once we have sufficient, i.e. several thousand, measurements, one can make pretty good estimates of the quality of measurements a buyer can expect. Aside from the practical utility, I think it would be an interesting study in the quality of low budget Chinese electronics manufacturing.

To collect the data I'll set up an email account for people to send their results to and write a program which will collect the data. I already have software which will compute the mean, standard deviation and the probability density function at each frequency. Then on a regular basis I'll rerun the statistical codes, plot and post the results.

It will take me a few days to develop the data collection routines as I'll need programs for multiple platforms. If a python program proves portable to Windows, Linux, Solaris and MacOS I'll use that, The TDR program uses that. So users have strong incentive to install python. And I figure it's about time I learned it.

At the moment the question is: 101 samples from 70 kHz to 900 MHz? 901 samples from 1 MHz to 900 MHz? or something different?

Have Fun!
Reg

Hi Reginald
Interesting approach but have you considered that the calibration kits used then should be from the same supplier else you will have a spread due to these are having different frequency characteristics and thus add an uncertainty not justifying the nanoVNA hardware investigation you are looking for. Just the short used can be with welded center pin and other with turned disk/center pin, which will introduce a couple of picosecond variation. Likewise for the open end cap there are many versions I have at least 4 different models where the one form Amphenol RF has a disk that close to the center conductor of the female adaptor on the nanoVNA that it lead to high delays and even often create a short. Other types have different mechanical structures, actually the best open is nothing used for frequencies up to 1GHz.
The load will definitely have frequency dependencies where both the resistance and inductive / capacitive component influence the frequency characteristic. There will as well be deviation from sample to sample and from brand to brand.
These comment not to kill your project but just to raise some concern
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Reginald Beardsley via Groups.Io
Sendt: 11. september 2019 00:21
Til: [email protected]
Emne: Re: [nanovna-users] Calibration data statistics

I'm suggesting something very basic. Accumulate enough data to make a meaningful statement of the mean and variance of S11 and S21 using the parts supplied with the unit by doing uncalibrated SOLT measurements with two open measurements. The bare SMA connector and the supplied open. Both ports for 5 cases, and if possible, a 6th case with both ports terminated with a 50 ohm load. No unknowns. Just poor quality "knowns".

Many of the people who buy these do not, and will not, have access to anything better. An Amphenol cal kit from SDR-Kits is roughly 1/2 the cost of the nanoVNA. The Rosenberger cal kit is 2x the price of a nanoVNA.

This is hobby science. The purpose is to provide a statistical estimate of measurement uncertainty. Fundamentally calibration consists of assuming that any deviation of the measured values from the theoretical values is device and fixture error. In this case we also have significant errors in the calibration standards.

Once we have sufficient, i.e. several thousand, measurements, one can make pretty good estimates of the quality of measurements a buyer can expect. Aside from the practical utility, I think it would be an interesting study in the quality of low budget Chinese electronics manufacturing.

To collect the data I'll set up an email account for people to send their results to and write a program which will collect the data. I already have software which will compute the mean, standard deviation and the probability density function at each frequency. Then on a regular basis I'll rerun the statistical codes, plot and post the results.

It will take me a few days to develop the data collection routines as I'll need programs for multiple platforms. If a python program proves portable to Windows, Linux, Solaris and MacOS I'll use that, The TDR program uses that. So users have strong incentive to install python. And I figure it's about time I learned it.

At the moment the question is: 101 samples from 70 kHz to 900 MHz? 901 samples from 1 MHz to 900 MHz? or something different?

Have Fun!
Reg

On Wed, 11 Sep 2019 at 12:49, Kurt Poulsen <kurt@...> wrote:

Hi Reginald
Interesting approach but have you considered that the calibration kits
used then should be from the same supplier else you will have a spread due
to these are having different frequency characteristics and thus add an
uncertainty not justifying the nanoVNA hardware investigation you are
looking for. Just the short used can be with welded center pin and other
with turned disk/center pin, which will introduce a couple of picosecond
variation. Likewise for the open end cap there are many versions I have at
least 4 different models where the one form Amphenol RF has a disk that
close to the center conductor of the female adaptor on the nanoVNA that it
lead to high delays and even often create a short.

Kurt and I don’t always agree on calibration issues???, but in this case
I agree with Kurt.

As I wrote yesterday. I had seen a paper where corrections were derived
when calibrating 3.5 mm connectors with 2.92 mm calibration kits and visa
versa. However corrections were not derived with SMA connectors due to the
variability among them.

Other types have different mechanical structures, actually the best open is

nothing used for frequencies up to 1GHz.

I would be surprised if that’s not the case *well* beyond 1 GHz.

I plotted the data the other day up to 900 MHz, showing added the open
supplied with my NanoVNA just increased the phase from the ideal 0
degrees, as it’s just adding unwanted capacitance.

The load will definitely have frequency dependencies where both the
resistance and inductive / capacitive component influence the frequency
characteristic. There will as well be deviation from sample to sample and
from brand to brand.

That’s why I select every load for calibration kits my company sells.
There’s too much variation among them

These comment not to kill your project but just to raise some concern
Kind regards
Kurt

I am only about 70% sure of Reg’s aim. But if it what I think it is, then
I believe the variation among the SMA bits will make it a bit pointless
trying to collect the data I believes he wants.

Unless good quality N connectors are used, the same issues will arise. In
particular the shape of the male pin N is poorly defined.

Dave


Dave.
--
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

David,

I'm trying to understand all this, and you are obvious an expert so I have a dummy question.
I added a simulation tool to my VNA software that uses a electrical model to calculate the input for the ADC of the nano VNA (the audio samples) and than the SW has to re-produce the values specified in the model. Its a great way to test and debug your SW.
When I use an "Open" and I apply a 1 cm increase of the length to the reference plane with a sweep from 1-900MHz and I compare it with a 1.3pF capacity over the "Open" I can not see the difference in the output of the VNA between addince 1cm and adding 1.3pF. Except a minute difference in the shape of the jX over frequency.
For Load the shift in reference plane does not change R and jX and for Close the capacitance does not make a difference but the shift in reference plane does.
So when you talk about "as it’s just adding unwanted capacitance.", isn't that unwanted capacity not the "same" as "adding one cm in distance to reference plane" when you are measuring Open and Load?
So for calibration, should the reference plane AND the stray capacitance (and inductance) be constant and ONLY the resistance change????

Hi David
Sure we have has some discussion ? and also some very good one.
I wrote 1GHz because that is the nanoVNA environment. I have measured to 8GHz on a 4 channel R&S VNA with auto calibrator and the Fairview SMA open cap, which is just an empty cylinder like the one from Pasternak, barely showed anything but a fraction of a degree in phase difference
Kind regards
Kurt

-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Dr. David Kirkby from Kirkby Microwave Ltd
Sendt: 11. september 2019 15:11
Til: [email protected]
Emne: Re: [nanovna-users] Calibration data statistics

On Wed, 11 Sep 2019 at 12:49, Kurt Poulsen <kurt@...> wrote:

Hi Reginald
Interesting approach but have you considered that the calibration kits
used then should be from the same supplier else you will have a spread
due to these are having different frequency characteristics and thus
add an uncertainty not justifying the nanoVNA hardware investigation
you are looking for. Just the short used can be with welded center pin
and other with turned disk/center pin, which will introduce a couple
of picosecond variation. Likewise for the open end cap there are many
versions I have at least 4 different models where the one form
Amphenol RF has a disk that close to the center conductor of the
female adaptor on the nanoVNA that it lead to high delays and even often create a short.

Kurt and I don’t always agree on calibration issues???, but in this case I agree with Kurt.

As I wrote yesterday. I had seen a paper where corrections were derived when calibrating 3.5 mm connectors with 2.92 mm calibration kits and visa versa. However corrections were not derived with SMA connectors due to the variability among them.

Other types have different mechanical structures, actually the best open is

nothing used for frequencies up to 1GHz.

I would be surprised if that’s not the case *well* beyond 1 GHz.

I plotted the data the other day up to 900 MHz, showing added the open supplied with my NanoVNA just increased the phase from the ideal 0 degrees, as it’s just adding unwanted capacitance.

The load will definitely have frequency dependencies where both the
resistance and inductive / capacitive component influence the
frequency characteristic. There will as well be deviation from sample
to sample and from brand to brand.

That’s why I select every load for calibration kits my company sells.
There’s too much variation among them

These comment not to kill your project but just to raise some concern
Kind regards Kurt

I am only about 70% sure of Reg’s aim. But if it what I think it is, then
I believe the variation among the SMA bits will make it a bit pointless
trying to collect the data I believes he wants.

Unless good quality N connectors are used, the same issues will arise. In
particular the shape of the male pin N is poorly defined.

Dave


Dave.
--
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

On Wed, 11 Sep 2019 at 15:08, <erik@...> wrote:

David,

I'm trying to understand all this, and you are obvious an expert

I would not say that.

so I have a dummy question.

I added a simulation tool to my VNA software that uses a electrical model
to calculate the input for the ADC of the nano VNA (the audio samples) and
than the SW has to re-produce the values specified in the model. Its a
great way to test and debug your SW.
When I use an "Open" and I apply a 1 cm increase of the length to the
reference plane with a sweep from 1-900MHz and I compare it with a 1.3pF
capacity over the "Open" I can not see the difference in the output of the
VNA between addince 1cm and adding 1.3pF. Except a minute difference in the
shape of the jX over frequency.
For Load the shift in reference plane does not change R and jX and for
Close the capacitance does not make a difference but the shift in reference
plane does.

I have not used my NanoVNA more than for an hour, and have not used your
software at all. Does your software allow one to specify

* Directivity of the bridge, as a vector
* Source match as a vector
* Frequency response of all components

I suspect that the answer is no, but rather the software assumes there are
no systematic errors.

If there are unknown systematic errors present, then it’s impossible to
predict what the VNA will display by adding a capacitor.

A port extension on a VNA is an analytical way of computing what will
happen if you shift the reference plane. That assumes that the systematic
errors have been removed too.

So I will be cautious of making assumptions about what will happen in the
NanoVNA, unless it has first been calibrated to remove the systematic
errors.

Dave.



--
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

My point is, if you buy a nanoVNA which comes with a cal kit, what accuracy can you reasonably expect from it? That is what I'd like to determine. It's a perfectly tractable problem given enough measurements. Ideally I'd like 100,000 or more, but a few thousand will probably have to suffice.

It is *not*. what can you do with a better cal kit. It is, what can you reasonably expect from what you bought? The full system error distribution, device and cal kit.

We know from fundamental analysis what the results should be for a *perfect* SOLT and a *perfect* VNA. Even though neither of these exist at *any* price. The calibration process attempts to correct for the errors imposed by the physical world.

A short and an open should have *exactly* the same magnitude, but be precisely 180 degrees out of phase.

The logistics of separating the nanoVNA errors from the cal kit errors is simply uneconomic. It would require shipping a precision cal kit to all participants. So all we can afford is the total system error of device and cal kit.

So far as I can tell, people are reading far more into the project I've proposed than is there.

All I want to do is quantify what the probable errors are for a random unit ordered online.

Reg

Hi Reg
Sensible argument, which will lead to a conclusion that all purchased are equally poor, as you will include poorly build as well the good quality ones. May I suggest you to try to sort your data after which brand is analyzed, as we know there are many manufacturers with various build quality. It will be unfair for those well build to be included in the huge mass of nanoVNA's. However that will be a hard job to trace which one the individual contributors have.
Anyway good luck
Kind regards
Kurt


-----Oprindelig meddelelse-----
Fra: [email protected] <[email protected]> P? vegne af Reginald Beardsley via Groups.Io
Sendt: 12. september 2019 04:38
Til: [email protected]
Emne: Re: [nanovna-users] Calibration data statistics

My point is, if you buy a nanoVNA which comes with a cal kit, what accuracy can you reasonably expect from it? That is what I'd like to determine. It's a perfectly tractable problem given enough measurements. Ideally I'd like 100,000 or more, but a few thousand will probably have to suffice.

It is *not*. what can you do with a better cal kit. It is, what can you reasonably expect from what you bought? The full system error distribution, device and cal kit.

We know from fundamental analysis what the results should be for a *perfect* SOLT and a *perfect* VNA. Even though neither of these exist at *any* price. The calibration process attempts to correct for the errors imposed by the physical world.

A short and an open should have *exactly* the same magnitude, but be precisely 180 degrees out of phase.

The logistics of separating the nanoVNA errors from the cal kit errors is simply uneconomic. It would require shipping a precision cal kit to all participants. So all we can afford is the total system error of device and cal kit.

So far as I can tell, people are reading far more into the project I've proposed than is there.

All I want to do is quantify what the probable errors are for a random unit ordered online.

Reg

On Tue, 10 Sep 2019 at 03:57, Reginald Beardsley via Groups.Io <pulaskite=
[email protected]> wrote:

I'd like to suggest is collecting calibration data and doing statistical
analyses.

You wrote sometime later that people are reading more into this than you
intended. I think you need to set out your aims more clearly, as I am
completely lost.??????

With calibration turned off, collect SOLT for 101 points over the 70 KHz to

900 MHz range using the console interface and post them to the list as
ASCII text: frequency, magnitude and phase.

What do you mean by “calibration turned off”?

In a professional VNA one can turn off the *error correction*. At that
point it is irrelevant whether the unit has been calibrated or not. Maybe
the NanoVNA has something similar, but calls it by another name.

What do you mean by “collect SOLT”

Having a large amount of such measurements would be immensely beneficial
to developing better FW and calibration for the devices. And as OEM QA.

To be useful FW and HW version and thru cable metadata is needed.

I assume you mean length, diameter, a part number in the unlikely event it
has one, colour, a description of the SMA connectors on the end etc. After
calibration, the effects of the cable should be removed, so these seem
largely irrelevant to me.

The only thing that seems relevant to me would be data on the female-female
thru, as currently the effect of its electrical length & loss can not be
allowed for during the calibration. But you don’t mention that.

I am just totally lost here.

The nanoVNA is an order of magnitude reduction in cost for a VNA. It's not

the equal of a VNWA or similar, but for many it is good enough and cheap
enough for price not to be an issue.

Yes. It also has the advantage of being entirely self contained. I have a
VNWA-3E here sitting unused. Not single measurement has been made with it
as I could not stand the tablet computer to use with it!

The biggest problems I see with it are

1) No support for the properties of the calibration kit.
2) No support for entering the delay of the thru used during calibration.
(This is related to #1 above.
3) The size of the screen.
4) Poor construction quality.

We've been handed the ball. Let's run with it!

Sure. I have a VNA that cost me $16,000 (used) and an N cal kit that cost
me $4000 (used), but I sure intend having some fun with this NanoVNA.

Have Fun!
Reg

Dave.

--

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