开云体育

I agree, return loss should be a positive number, and I can remember endless debates in our engineering-team about it, but when you tune filters on a Network analyzer, it is nice to have the insertion loss (S12) and return loss (S11) going the other way on a single screen. My memory is not too good after having a brain tumor and a couple of strokes, but I seem to recall the HP network analyzers (HP8712???) also showing S11 as a negative number, whilst S21 as a positive number. My memory is a bit fuzzy, bit I can remember the one graph, S12 going up (less negative), and S11 going more negative on the same screen, making it easy to tune. Whether it was the default mode for the HP8712 or whether it was a custom-mode, I can’t remember.

Hi,
For filters the steepness of the skirts is important apart from the 3dB bandwidth.
-6DB and -80 dB are a real measure for the skirts!
However ,many filters ( or measuring conditions ) are not attaining the -80dB , so -60dB will have to do.
This is OK as long as you realise or/and that this is CLEARLY mentioned/stated.
I have seen many tables with comparisons of filters and their steepness.
Steepness was then 80/3 or 60/3 dB value.
Older engineers told me that this was better than using the ratio of the frequency of the -3 and -60 dB points because the difference between the types of filters( number of poles/crystal,L-C;etc/etc.) could be better evaluated.
I know of discussions because of mixing between -60 and -80 dB values( don't compare two sorts of beasts)

Jan ON4MMW

The current firmware allows dumping the internal calibration table with the data command
data [0-6]
where
0: S11
1: S21
and
2 /* error term directivity */
3 /* error term source match */
4 /* error term refrection tracking */
5 /* error term transmission tracking */
6 /* error term isolation */
but it is currently NOT possible to upload calibration tables into the nanoVNA

a link about measuring filters

In my own experience coming from a R&D environment, we would specify filters with the -3dB and -60dB (theoretical) points, but for production and customer tests, AND specifications, we would only test the -3dB point and the -45 dB points, as those points were achievable by most equipment. The -60dB point was considered to be theoretical as very few customers could measure it.

Hello Run,

I like to use your program and I am very happy with it.
You try to realize feasible expansion requests during the program development.
And all for free and for the three main operating systems. That's a great achievement.
Thanks a lot for this.
The criticism of someone sitting on a high horse should not impress you.

73, Rudi DL5FA

On Fri, 4 Oct 2019 at 13:38, Martin via Groups.Io <martin_ehrenfried=
[email protected]> wrote:

i All,

Owen Duffy has recently posted a note about the Nano VNA on his blog.



He makes a few points about the accuracy of calculated values with both
the Nano VNA and Saver in particular.

Whilst agreeing on his return loss statement, the blog is written in a
childish condescending way. If anyone wants to forward him the post I made
about how the HP LCR meter can display things, they are welcome to, but I
can't be bothered to communicate with people like that.

Dave

If we first deal with the issue of Return Loss, which is probably the most
problematic par, and has certainly caught me out on several occasions.

Just about every instrument I can remember using, has RL shown as a
negative curve, even if the units themselves are positive. This is handy if
for example you are tuning a filter, as you can see the insertion gain on
the uppermost trace and the RL loss on the lower one without them
overlapping. It also matches the convention of SWR plots and when
measuring the RL of cables it matches the convention of more attenuation
being negative.

However Owen makes the point that negative loss is actually gain (double
negative) and vice versa, and the existing conventions do indeed lead to
confusion and mistakes being made. Maybe return loss should really be
called return gain, and then everyone would be happy (well maybe - but this
is not a serious suggestion).

However if we put this to one side, there is still the issue of how the
values are being calculated, and if they are in fact correct. If not then I
think this should be investigated in more detail and fixed, as there would
seem to be an opportunity to do this before it propagates further.

Regards,

Martin - G8JNJ

--
Dr David Kirkby Ph.D C.Eng MIET
Kirkby Microwave Ltd
Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, CHELMSFORD,
Essex, CM3 6DT, United Kingdom.
Registered in England and Wales as company number 08914892

Tel 01621-680100 / +44 1621-680100

Rune,
My school days are far behind me, but I recall that different filter designs fall off in certain dB/decades (i.e. 20 dB/decade) so if you were trying to extrapolate it would definitely help to know what type of filter is being measure and the number of poles. Otherwise you will have to do something empirical and look at trend in the dataset and do an extrapolation based on the trend from say -3 dB or - 6 dB to -40 dB which should well out of the NanoVNA's noise floor.

I'm sure members with more experience in this area will weigh in.

Herb

30 : our final report 1

29 : ann : our nanovna will be evaluated tonight :
/g/nanovna-users/message/3867

hello all,

this is our final report on the comparison of our
[nanovna] and [vna] in terms of frequency

1 : for nominal values without uncertainty :

red : nanovna - blue : vna

r :
x :

after all that said, accept, please, our humble - but sincere - congratulations

sincerely yours,

gin&pez@arg

30

Hi Herb,
my code thus far does indeed try to extrapolate what the -60 dB point would
be. Only I'm nowhere near good enough at making calculations like this, so
I'm stumbling a little. I know what the -6 dB point is, and I know
(roughly) how many dB per octave or decade the filter drops by. I think I
then need to do something magic involving logarithms, probably something
really simple - but I haven't sat down and figured it out yet.

Suggestions welcome if it's obvious to you, the reader. :-)

--
Rune / 5Q5R

On Fri, 4 Oct 2019 at 16:32, Oristo <ormpoa@...> wrote:

its become the consensus that return loss is a negative number

IMO, "return loss" is a poor term.

But one used very extensively. Personally, I don't have a problem with it.

Arguably, >>any<< return is loss, when sending power is the goal. In which

case,
"return loss" might be power sent (since lost from return)..

Not always. There are one-port amplifiers, such devices as Gunn diodes and
IMPATT diodes, but they are quite rare. For them, the signal returned is
greater than that sent to them. They work on the principle of negative
dynamic resistance. That means increasing the voltage, decreases the
current. They still have a positive resistance though. A device might draw
100 mA at 7 V, but 90 mA at 8 V. The voltage has increased, but the current
goes down.

I know what my opinion on the subject is (return loss is positive for any
passive device), but I am not going to try to force it on anyone else.

I'm an atheist, but I would never try to convince anyone of any religion
that their belief is silly. It is a waste of my time.

--
Dr David Kirkby Ph.D C.Eng MIET
Kirkby Microwave Ltd
Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, CHELMSFORD,
Essex, CM3 6DT, United Kingdom.
Registered in England and Wales as company number 08914892

Tel 01621-680100 / +44 1621-680100

Hi Martin,
thank you very much for your appreciation! :-)

I don't know Owen, so I couldn't possibly comment on his experiences; nor
will I speculate on where his heart is. I only know what impact comments
like his can have on the motivation of a developer.

David suggested a number of different readings in a separate post that I
think one of his pieces of hardware shows. All those could be interesting
to add, perhaps to a pop-out display in order to fit them all ;-)

The user definable SWR markers have been a request, and I think I'll put
them in for both Smith, SWR - and of course, "return loss" ;-)

--
Rune / 5Q5R

On Fri, 4 Oct 2019 at 15:11, Martin via Groups.Io <martin_ehrenfried=
[email protected]> wrote:

Hi Rune,

Understood. I really like Saver and very much appreciate what you have
achieved so far, especially in such a short space of time, and I'm very
grateful for your effort and dedication.

A lot of Owen's frustration is due to problems with other VNA's and
antenna analysers that have some serious issues with calculated values that
never seem to get fixed, and the negative response he has had from the
manufacturers when these have been flagged up, hence his reluctance to
engage directly.

I agree that amateur conventions often differ from those used in the
professional world, and that perhaps some of Owen's comments are merely a
reflection (no pun intended) of that conflict. However is heart is in the
right place when trying to point out these issues, even if perhaps his
methodology is not to your liking.

The AIM VNA shows both parallel and series values, so maybe it would be
useful to include those in addition to conductance & admittance and then
it's not necessary to resort to external calculators and spreadsheets ?

Finally - would it be possible to add some user definable SWR marker lines
to the Smith and SWR plots ?

Regards,

Martin - G8JNJ

Mike,
Google on filter bandwidth and the majority of the results will come up as the -3 dB point. Not saying the -6 dB and -60 db points are not useful info, just that the -3dB point is the common reference.

One other point is that -60 dB is in the dirt for portions of the nanoVNA measurement range. Rune would have to extrapolate the -60 dB point in that range which lends itself to possible inaccurate assumptions of the filter's behavior.

Herb

Lower frequency limit: will be unchanged because the si5351 is still present and is switched in below 140MHz.

Larger screen: we are going to use the plug-in style of ILI9341 display rather than the bare FPC version because this board has components on both sides, so the 3.5 inch ILI9341 display is drop in compatible (provided you buy one with the correct pinout). Size beyond 3.5inch is not currently doable because 3.5inch is the largest SPI LCD size available. Larger displays (for example on the NanoVNA-F) use a parallel interface with far more pins which will require a larger MCU, adding many dollars of cost.

MAX2871: unfortunately this costs about 30 times the ADF4350, so isn't doable for the base version.

IAM-81008: that's a nice part, cheaper than the AD8342 and less power consumption. We might evaluate it and see if it will give acceptable performance. The IIP3 is much lower though, which means it has to be operated at lower signal levels to avoid bad linearity.

Open source: the schematic will be posted as soon as the product is generally available.

In the longer term we may introduce a higher end VNA. However the bigger priority there is full two port support - measuring all 4 S parameters at once. This is more aimed at lab use, so we may just remove the screen altogether (it's a big cost adder and a nuisance for shipping and firmware development) and have only a USB and wireless user interface.

Martin,
You might also try using the console commands via a terminal to accomplish saving and loading configurations. If it can be done via a terminal, then you can provide that feedback to Rune and it will make it easier for him to incorporate that function into his application. Larry has just added an updated list of commands to the "Files" section.

Herb

its become the consensus that return loss is a negative number

IMO, "return loss" is a poor term.
Arguably, >>any<< return is loss, when sending power is the goal. In which case,
"return loss" might be power sent (since lost from return)..

Perhaps the nanoVNA-F could be added?

The document points to /g/nanovna-f

Thanks! Rather a pity that the two couldn't be combined.

I appreciate that some folks here seem interested in both.
Many messages with potentially useful content already here
lack context and often stray far from nominal topic,
requiring "reverse engineering" to exploit.

I focus on sorting descriptions of function for only NanoVNA-H
e.g. among firmware flavors.
For other examples, Windows 7 and Windows 10 user experience
differences are IMO poorly addressed by current documentation, and
I feel badly about not updating interactive menu tree for newer firmware..

For my own relatively trivial few contributions,
version control and documentation update coordination
provoke quite a lot of what is described for ducks as
"furiously paddling beneath the surface",
but I am only a poor and slow typist.

I read a recent post that said that, among hams, its become the consensus that return loss is a negative number. I respectfully disagree.



Regarding return loss, I understand some think its a negative number, some a positive. I am in the later camp. Why? When I was a young engineer I had the great fortune, in the early 1980s, to work for HP. Arguably the premier test and measurement company of its time. And equally arguably a leading if not THE leading microwave and rf company.



As new sales people, we were sent to 3, 3 week training sessions to learn the technology and the products so we might interact with our engineering customers in a technically sound way.



One thing HP taught, as I've said in a much earlier post, was that return loss was always a positive number. As someone on this list pointed out, negative loss is gain. And we KNOW that no return loss measurement, of a passive device, can exhibit gain.?



With respect, there is no gray area here. Return loss, is always a positive (or, perhaps better said, a non- signed) number.?


Even if you are at a hobby level of involvement, why not use the proper, and correct, terminology? If you're a private pilot, you are expected/ required to use the proper terminology. Same if you are an amateur road racer. Or "fill in your favorite" hobby.?













Sent using

If you ever bought one of the popular starter kits for microcontrollers like Arduino, it probably contains some suitable switches.

I bought a bag of assorted varieties of these switches from Banggood so I could try out the various form factors, especially the length of the button: In that assortment they are about 3 cents per switch. If you need a bunch of a single type, you can buy bigger bags of them and they are even cheaper.

There has been some discussion about the best way to represent data
measured as impedance, into an equivalent circuit. The most comprehensive
of these I have seen, which my company owns, is my obsolte HP 4284A
precision LCR meter.



which is *not* a VNA, but measures the impedance of a device using a 4-wire
Kelvin measurement. This means everything shown here could conceivably be
displayed.

Someone *might* want to implement all or a subset of these. *I'm not
suggesting they are all implemented - I am just listing them for possible
discussion. *I suspect my HP 4285A LCR meter (75 kHz to 30 MHz) does the
same, but I have not checked.

*HP 4284A 1 MHz LCR meter. *
Cp-D - a parallel capacitance Cp, with a dissipation fact D
Cp-Q - a parallel capacitance Cp, with a Q of Q
Cp-G - a parallel capacitance Cp, with a conductance G
Cp-B - a parallel capacitance Cp, with a susceptance B
Cp-Rp - a parallel combination of a capacitance Cp, and resistance Rp
Cs-D - a series capacitance Cs, with dissipation factor Q
Cs-Q - a series capacitance Cs, with a Q of Q
Cs-Rs - a series combination of capacitance Cs and resistance Rs
Lp-D - a parallel inductor, with a dissipation factor D
Lp-Q - a parallel inductor, with a Q of Q
Lp-G - a parallel inductor, with a conductance of G measured in siemens
Lp-Rp - a parallel combination of an inductance Lp and resistance Rp.
Ls-Q - a series inductor Ls, with a Q of Q
Ls-Rs - a series inductor Ls with resistance Rs

*R - X - I don't need to explain what that is. *
Z-theta (degrees) - The magnitude and phase of an impedance expressed in
degrees.
Z-theta (rad) - The magnitude and phase of an impedance expressed in
radians.
G-B - G is the conductance measured in siemens (admittance Y = G + j B)
Y-theta (degrees) the magnitude of the admittance Y, and phase angle
expressed in degrees
Y-theta (rad) the magnitude of the admittance Y, and phase angle expressed
in radians

--
Dr David Kirkby Ph.D C.Eng MIET
Kirkby Microwave Ltd
Registered office: Stokes Hall Lodge, Burnham Rd, Althorne, CHELMSFORD,
Essex, CM3 6DT, United Kingdom.
Registered in England and Wales as company number 08914892

Tel 01621-680100 / +44 1621-680100